Interrupting.. but i have a question... can i use TIP3055(90w) & MJE2955(75w) for small 20-25w complimentary pair amplifier experiments? Of course within SOA & with perfect thermal management. But still.. does it effect anything? 
Manufacturers have been known to use 30 watt Jap TO-220’s or TIP TO-220’s on sufficiently small amplifiers. You often see the “8 ohm minimum load” warnings on TO-220 amplifies if they push the rail voltages too high. Biggest problem is the bulk heat removal from the package - 20 watts EACH is much as you can safely do real world - and that’s on a big heat sink. You want to be dissipating in the 10 To 12 watt range, per package, max - on average. Just keep it to about 48V rail to rail and you won’t have thermal problems. That will also keep the peak dissipations well under 75W, even with reactive loads.
TIP version, MJE… doesn’t really matter. The case to sink thermal resistance is the same bottleneck either way.
TIP version, MJE… doesn’t really matter. The case to sink thermal resistance is the same bottleneck either way.
Ok, but i think you missed my point or maybe i failed to explain. I was trying to say that can i use 90w NPN with 75w PNP for the same amplifier?
@anatech - Myself I work with ATE systems (mostly NGATS and other Mil systems) as a day job - there we have several of the Fluke 5720A calibrators and you can't walk without tripping on a 3458A - all to adjust systems to, at best, +- 0.5% Gotta love the military way of doing things. Anyway, the actual metrology is my hobby. Well I guess more accurately precision analog, but pretty much the same thing usually. My actually use currently for power devices is to use in several active load applications I'm working on. My first thought was to use some of the Linear series from IXYS but it would take a water cooling system to really utilize their capabilities. I could use standard MOSFETS suitably de-rated to comply with their SOA but then I remembered I had several hundred perfectly good power transistors already along with heatsinks already drilled to accommodate them. Perfect matching isn't necessary for the application as they will be individually driven but close is better than not. This is the type of application that I am familiar with but not directly applicable to audio I guessed - and the info provided here has proved that to be correct.
But i have a plan for that.. i'll mount MJE directly to the heatsink & will use insulator only for TIP part.
@anatech - On the subject of personal bench equipment, I currently use a DMM7510 which I love, but my 34401A is there along side it and will never be retired if I can help it - absolutely rock solid meter and for the last fifteen years I've had it has never drifted out of spec. The same can be said for my 3456A - it's a beast but rock solid. Personally, I don't like the 3457a - good, solid meter but I don't like the case and hate the display. I have quite a few of the 3456As in storage and several 3455As as well - keep telling myself to refurb and adjust to sell but haven't managed to convince myself to do that yet. Don't have (or have any plans to get) a 3458A but, dang, they are a nice piece of kit! Designed 30+ years ago and still unrivaled - I've learned a lot studying how they managed to do it; truly amazing engineering.
Then the TIP part will end up running hotter. An insulator doesn’t just hurt the TO-220 thermal path a little - it hurts it a LOT. It’s why you limit Pdiss to 20W. 40 is possible without an insulator If you use good thermal compound in the proper amount. But no one ever does.
My solution for TO-220 amps is to use CFPs. Individual heat sinks on the outputs and no insulator.
My solution for TO-220 amps is to use CFPs. Individual heat sinks on the outputs and no insulator.
Don't they have the same case? Then they will get equally hot.
The spec doesn't tell them how hot they get, the power supply voltage and bias current determines the dissipation, and if they have the same case, they get equally hot.
Jan
No, TIP3055 is bigger TO-247 package, MJE2955 is TO-220. That's why i'll not use insulators for MJE part.
Insulator on TO-247 isn’t as much of a problem because of the much larger surface area AND the torque on the screw actually puts pressure where it is needed due to the more rigid package. I thought both were TO-220’s. Depends whose TIP.
If you’re building a singe supply EF2 amp with an output capacitor you could actually still ground the heat sink.
If you’re building a singe supply EF2 amp with an output capacitor you could actually still ground the heat sink.
Se what Jan say is correct. However, the package type will determine what heat transfer (thermal conductivity) is relevant, just look at the spec sheet. Watts remains watts, there is no getting away from it.
Yes. You'd want to look at the thermal resistance. The data sheet normally shows T(junction to case) Tjc, in degrees per watt.
So for instance with a 10W dissipation and a Tjc of 5degrees/watt, the junction will be 50 degrees hotter than the case.
So if you want to keep the junction below 100 degrees, you need a case below 50 degrees, and that determines the heatsink you need.
All simple algebra ;-)
Jan
So for instance with a 10W dissipation and a Tjc of 5degrees/watt, the junction will be 50 degrees hotter than the case.
So if you want to keep the junction below 100 degrees, you need a case below 50 degrees, and that determines the heatsink you need.
All simple algebra ;-)
Jan
Hi Hal,
Yup, I have a 3456A on the bench as well. The 3457A has a 300VDC limit, so I can't use it for some things. It is very accurate though. The display doesn't degrade like my 34401A's did. I expect the OLED (current line of meters) displays to die also, not happy with that choice!
Back to transistors ...
Yes, TO-220 parts are also sensitive to tab warping which can crack the die loose or break the package seal. I have seen mis-matched output devices and packages (in repairs - believe it or not). They can function, but this is not recommended. Plastic cased parts are cheap enough, just buy the complementary pairs and be done with it.
Running any part close to thermal limits isn't wise. This is one area where you need to be conservative. The failure rate doubles for each 10 °C rise in temperature.
Yup, I have a 3456A on the bench as well. The 3457A has a 300VDC limit, so I can't use it for some things. It is very accurate though. The display doesn't degrade like my 34401A's did. I expect the OLED (current line of meters) displays to die also, not happy with that choice!
Back to transistors ...
Yes, TO-220 parts are also sensitive to tab warping which can crack the die loose or break the package seal. I have seen mis-matched output devices and packages (in repairs - believe it or not). They can function, but this is not recommended. Plastic cased parts are cheap enough, just buy the complementary pairs and be done with it.
Running any part close to thermal limits isn't wise. This is one area where you need to be conservative. The failure rate doubles for each 10 °C rise in temperature.
Biggest problem is case to sink thermal resistance for a TO-220. Using the mounting hole and an insulator, it’s about 3 C per watt. It sucks. To get it better you have to dispense with the insulator, use a bar clamp across the package body, or both
Industry holds themselves to a different standard tha what they sell to consumers.
When something breaks/burns out it costs them money.
When something breaks/burns out it costs them money.
But then, industry pays well for that standard. As with everything else, you get what you pay for - and true, quantifiable reliability isn't cheap.
I'd have to disagree with you Hal. The consumer pays- through the nose, and does not get what they pay for. Not anymore. There is no such thing as "durable goods" as a classification. Not really. No readily accessible service information any more (right to repair is a joke). Parts availability is nearly non-existent. Designs are engineered to fail, sending products in record numbers to the land fill. As a world, we cannot afford it. Never mind the energy and pollution transporting that crap.
Industry doesn't pay high amounts to get reliable products. They cannot afford to use the junk released to the consumer market. We cannot afford that junk either to be honest. Industrial standards are slipping, but it used to be that consumer equipment was designed in a similar way. You should see the unreliable junk sold to the telecom industry today!
All markets are geared to giving their customer less at the highest price the market will bear. This will break at some point.
Hi wg_ski,
Don't kid yourself. The cost to consumers is enormous! We just haven't added it all up yet. Then look at lost jobs in all sectors.
I guess greed tricked us into voting with our wallets. We are not going to like where this road leads, some have seen it for years.
Industry doesn't pay high amounts to get reliable products. They cannot afford to use the junk released to the consumer market. We cannot afford that junk either to be honest. Industrial standards are slipping, but it used to be that consumer equipment was designed in a similar way. You should see the unreliable junk sold to the telecom industry today!
All markets are geared to giving their customer less at the highest price the market will bear. This will break at some point.
Hi wg_ski,
Don't kid yourself. The cost to consumers is enormous! We just haven't added it all up yet. Then look at lost jobs in all sectors.
I guess greed tricked us into voting with our wallets. We are not going to like where this road leads, some have seen it for years.