peranders said:
Yes, the datasheet does say that.
And you are correct about 4 to 6 under normal condition.
All must note the conditions being specified. With pc mount, the pad area and the lead length are crucial to the junction temperature and the derating.
I had 3 amp diodes running 21 amps DC with the junction at 150C (measured) and the leads at zero length and 25 C..But I don't recommend it. I was performing thermal characterization of the silicon/braze integrity.
Cheers, John
many years ago, there were 5 watt zener diodes used in consumer equipment that were placed in circuits requiring a 10 watt zener. the company (philips, i think) got around the problem by soldering a heat sink to the anode lead of the diode. it still wasn't very reliable, but at least the zener lasted about long enough for the warranty to expire....
crippledchicken said:
Why not go extreme overkill.......
http://www.fairchildsemi.com/ds/RU/RURG8060.pdf
I'm building a bridge with these right now actually.
just went into Jaycar (aussie electronics shop) to buy 12 x 47nF X2 250v mains caps to use on my rectifiers. Not in stock, "we can get 5 in". I'm wondering why he can only get 5 "I need 12" I say. Then he looks at me suspiciously, looks around a couple of times and whispers "I can back order you some".
blimey, I'm not trying to bypass a sub station. I just want my mid-highs to sound nice and smooth
blimey, I'm not trying to bypass a sub station. I just want my mid-highs to sound nice and smooth
there seems to be two types of engineers in demand in the consumer electronics industry, and they both keep my end of the business (repair) busy. the first one is the one who does anything to save a few cents, and the second one designs bizarre circuits that replace something simple and reliable, just to prove that they can. they're at opposite ends of the spectrum, but neither of them design a reliable product. i call them "Mr Cheapo, and Mr Conundrum". i'll admit i've had a few ideas for conundrums http://www.diyaudio.com/forums/showthread.php?threadid=121685&goto=newpost
and i've done quite a bit of penny pinching (but always test for reliability first before saying "we can save money by doing this"). my "design philosophy" is "do it simple, and do it well". notice i said "simple" and not "cheap". i may come up with a cost-effective solution to a problem, but i'll use overkill on component specs. i don't want to have something come back to bite me.
an example of Mr Cheapo's work is the use of a voltage multipler circuit in the Yamaha AV-RX661 receiver that feeds the -30V rail for the display tube. the circuit uses 4 100mA diodes and 4 electrolytic caps (85 degree temp rating). the caps are right next to a regulator heat sink and the diodes can't handle the charging current for very long. i have been replacing them with 105 degree caps and 1A diodes. Mr Cheapo probably saved a few fractions of a cent on each component, but created a high failure item. Mr Conundrum isn't satisfied with protection circuits that simply shut off the speaker relay when a fault is detected, so he feeds the detection circuitry back to a microprocessor that cuts off the main power immediately, and the protection circuits are more complex than the amp itself, so more goes wrong with them than with the amp.
and i've done quite a bit of penny pinching (but always test for reliability first before saying "we can save money by doing this"). my "design philosophy" is "do it simple, and do it well". notice i said "simple" and not "cheap". i may come up with a cost-effective solution to a problem, but i'll use overkill on component specs. i don't want to have something come back to bite me.
an example of Mr Cheapo's work is the use of a voltage multipler circuit in the Yamaha AV-RX661 receiver that feeds the -30V rail for the display tube. the circuit uses 4 100mA diodes and 4 electrolytic caps (85 degree temp rating). the caps are right next to a regulator heat sink and the diodes can't handle the charging current for very long. i have been replacing them with 105 degree caps and 1A diodes. Mr Cheapo probably saved a few fractions of a cent on each component, but created a high failure item. Mr Conundrum isn't satisfied with protection circuits that simply shut off the speaker relay when a fault is detected, so he feeds the detection circuitry back to a microprocessor that cuts off the main power immediately, and the protection circuits are more complex than the amp itself, so more goes wrong with them than with the amp.
How very true Some design is so poor that even top spec parts can't save them. Like the 12volt rail in a TV that is derived by a series resistor, think it was about 6.8ohms and a 12 v zener to ground from a line derived (16 volt I think) rail. Any guesses for what happens if the HT is a couple of percent high. And why do consumers have to pay when the manufacturer brings out a "reliability upgrade kit" Reliability kit ? some are just a complete rebuild of every piece of silicon in the PSU. I am sure some manufacturers throw a bag of bits in the air and see how they land. That's it, the latest SMPS with ony 30 SMD transistors and as many SMD diodes. And what is it with Philips and zeners.
Or the 100k pot thats fed by a 221.425795 kohm resistor. Would 220k not work when it's in series with a 100k pot.
Glad I am out of that side now.
Happy days,
Karl
Edit, Nice link by the way.
Some design is so poor that even top spec parts can't save them. Like the 12volt rail in a TV that is derived by a series resistor, think it was about 6.8ohms and a 12 v zener to ground from a line derived (16 volt I think) rail. Any guesses for what happens if the HT is a couple of percent high. And why do consumers have to pay when the manufacturer brings out a "reliability upgrade kit" Reliability kit ? some are just a complete rebuild of every piece of silicon in the PSU. I am sure some manufacturers throw a bag of bits in the air and see how they land. That's it, the latest SMPS with ony 30 SMD transistors and as many SMD diodes. And what is it with Philips and zeners.Or the 100k pot thats fed by a 221.425795 kohm resistor. Would 220k not work when it's in series with a 100k pot.
Glad I am out of that side now.
Happy days,
Karl
Edit, Nice link by the way.
yeah, the philips "10 watt" zener was a terrible way to get 12V from what IIRC was a 65V rail. i think the last time i saw one of those come back for a repeat repair, i built a series pass regulator with a zener and a bjt and the circuit never failed again. a lot of VGA monitors used a TDA1170 for a vertical amp. the compaq 460 monitor had a lot of trouble with those. the TDA1170 was never designed for 14" monitors (the original data sheets said it was for 5" 10" and 12" monitors), and compaq's particular design had the TDA running 3A spikes for 2 seconds every time you turned the monitor on. i had so many of them fail that i built a device that ramped the PS rail up over the period of 5 seconds to keep the TDA chips from self destructing. it was really simple, a couple of resistors, a cap and an IRF510 mosfet. the device connected in line with the 24Vrail for the TDA chip and limited the startup current enough that the TDA chip never failed again. yup, to save a few pennies, Mr Cheapo will run components at the bleeding edge of their ratings, and pray it stays together long enough for the warranty to expire.
Speaking of stupid designs in circuitry..........
When I was about 17, I worked on an older High-Power Fisher Home Stereo Receiver.
Seemed like an awesome amp, ..........except a few design shortcuts. It had great sound.
This amp was 150W/ch @ 8 ohms, had unloaded rails of +/- 70V, and used 150W 15A TO-264 Toshiba transistors (don't remember which part#)
Amp had a huge heavy transformer, big heatsink, but only had average power caps, and bridge rectifier.
1. The amp used a combined transistor for the LTP (looked like 6 pin IC) for some would be harder to replace.
2. The amp itself used little TO-126 transistors for VAS/CCS and drivers and NO heatsinks. The normally beige PCB was really dark brown around all those transistors from years of heat. They ran so hot you could barely touch them! Being +/-70V rails, with no Heatsinks, I can see why! Just heatsinking the transistors it already had would have been a major improvment.
3. For such a high power output, it only used 1 pair of transistors per channel!!! When guy had one channel go out, he hooked up both speakers into the one remaining channel, and it blew too. IMO the amp should have had 2 pairs per channel.
That's why when I DIY, I overbuild my circuits to make more reliable, because seeing stuff like this. Just more or better transistors, heatsinking, and being thoughtful of thermal limits makes a design better.
When I was about 17, I worked on an older High-Power Fisher Home Stereo Receiver.
Seemed like an awesome amp, ..........except a few design shortcuts. It had great sound.
This amp was 150W/ch @ 8 ohms, had unloaded rails of +/- 70V, and used 150W 15A TO-264 Toshiba transistors (don't remember which part#)
Amp had a huge heavy transformer, big heatsink, but only had average power caps, and bridge rectifier.
1. The amp used a combined transistor for the LTP (looked like 6 pin IC) for some would be harder to replace.
2. The amp itself used little TO-126 transistors for VAS/CCS and drivers and NO heatsinks. The normally beige PCB was really dark brown around all those transistors from years of heat. They ran so hot you could barely touch them! Being +/-70V rails, with no Heatsinks, I can see why! Just heatsinking the transistors it already had would have been a major improvment.
3. For such a high power output, it only used 1 pair of transistors per channel!!! When guy had one channel go out, he hooked up both speakers into the one remaining channel, and it blew too. IMO the amp should have had 2 pairs per channel.
That's why when I DIY, I overbuild my circuits to make more reliable, because seeing stuff like this. Just more or better transistors, heatsinking, and being thoughtful of thermal limits makes a design better.
Absolutely A darkened PCB, I asked manufacturer about this on early satellite set top box (very early actually, linear PSU not SMPS) Answer, this has been taken into account in the design. The material is of a type meant to discolour with heat. Only problem, the bridge rectifiers ran so hot the discoloured print actually carbonised and the diodes got that hot the solder was melting.
Think we better start another thread for all the disasters we have seen, going a bit off the topic I fear.
Karl
A darkened PCB, I asked manufacturer about this on early satellite set top box (very early actually, linear PSU not SMPS) Answer, this has been taken into account in the design. The material is of a type meant to discolour with heat. Only problem, the bridge rectifiers ran so hot the discoloured print actually carbonised and the diodes got that hot the solder was melting.Think we better start another thread for all the disasters we have seen, going a bit off the topic I fear.
Karl
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.