Is there a way to load test outputs out of circuit on the bench? With all the counterfeit transistors out there I'd like to build a load tester and run them at the max SOA and see if they hold up or fail. It seems a better way then to damage our projects. Does anyone have a circut design ? I have been ripped off in the past so this is why I ask.
This is not a simple circuit. The voltage and current are high and the test has to be pulsed. It is easiest to buy only from the top distributors.
Ed
Ed
I agree, why compromise your project with fakes. Properly load testing them will likely make most of them them fail so what's next?
Hugo
Hugo
From what I have been reading because most of the better transistors have been discontinued the ones being offered as genuine are usually fakes. Even the suppliers that usually have the genuine ones fakes have slipped through. Trust me if you know of any suppliers with inventory I am ready to purchase from them. It is a sad day when diy's are getting scammed by these criminals. I understand why manufacturers discontinued most of these products. The demand wasn't there anymore.
I would think that the big suppliers won't risk their reputation in the industry.Even the suppliers that usually have the genuine ones fakes have slipped through.
Farnell, Newark, RS online, Digikey, Mouser to name a few are very trustworthy.
Hugo
Is there a way to load test outputs out of circuit on the bench? With all the counterfeit transistors out there I'd like to build a load tester and run them at the max SOA and see if they hold up or fail. It seems a better way than to damage our projects. Does anyone have a circuit design? I have been ripped off in the past so this is why I ask.
That would be an interesting circuit. 👍
It might've prevented a situation that I found myself in, a few days ago.
I found that an emitter-follower-based, regulated DC PS was faulty - the transistor (a KSC3503) delivered the correct DC output voltage (24v, set by a zener) when it wasn't under load - but as soon as 20ma of current was drawn from it ... the output voltage collapsed to 12v! 😵
I had checked the transistor with an Atlas DCA75, before soldering it in, to see that it was working ... but that doesn't test it under load. 🙁
Есть ли способ нагрузить тестовые выходы вне цепи на стенде? Со всеми контрафактными транзисторами я хотел бы создать нагрузочный тестер и запустить их на максимальной SOA и посмотреть, выдержат они испытание или потерпят неудачу. Кажется, что это лучший способ навредить нашим проектам. У кого-нибудь есть схемотехника? Меня обманывали в прошлом, поэтому я спрашиваю.
Attachments
Easies way is to rig it as a current source. I use a 50V 3 amp test for most old TO-3’s. Will work to screen 2SC5200’s or other flatpacks too. Setup can be very simple and doesn’t need precision. 65V on the collector, 15 on the base, and a 4.7 ohm resistor (4 in series parallel you need a 50 watter) in the emitter. The test does NOT need to be pulsed at all - just at a controlled temperature. A finned heat sink with part of the fins in an ice bath works. That will keep Tc at 25C or less. Scale the emitter resistor for smaller transistors. Use a bench supply for the base. A 48 volt transformer and rectifier is fine for the collector. Doesn’t even need to be filtered - just let the peaks hit 65-68 volts. Fuse (or current limit) the base at 1A, and fuse the collector at 7. If it blows it blows. Then you knows.
Screening vceo is easy. I start with a 340VDC power supply and a 100k resistor. Simply measure the voltage you get, and stay well under. Can sort unmarked zeners and unknown transistors (or ones the marking has worn off) that way too so it’s a good idea to leave that rig together too.
The problem with smaller transistors (ie 2SD699, 2SC3503) is that fakes will normally PASS the manufacturer’s published SOA. Look at their data sheets, the power handling isn’t wonderful. If you need to screen these types you need a way to test speed because thats where the differences are going to show up.
Screening vceo is easy. I start with a 340VDC power supply and a 100k resistor. Simply measure the voltage you get, and stay well under. Can sort unmarked zeners and unknown transistors (or ones the marking has worn off) that way too so it’s a good idea to leave that rig together too.
The problem with smaller transistors (ie 2SD699, 2SC3503) is that fakes will normally PASS the manufacturer’s published SOA. Look at their data sheets, the power handling isn’t wonderful. If you need to screen these types you need a way to test speed because thats where the differences are going to show up.
Use a simple diy power amplifier and a heavy load.
Use a function generator with variable frequency (sweep) as input signal.
Adjust attention for max amplifier output.
Place fuses in each power supply polarity.
Stay away!💥
Use a function generator with variable frequency (sweep) as input signal.
Adjust attention for max amplifier output.
Place fuses in each power supply polarity.
Stay away!💥
Way to go: actual testing at real world conditions.Easies way is to rig it as a current source. I use a 50V 3 amp test for most old TO-3’s. Will work to screen 2SC5200’s or other flatpacks too. Setup can be very simple and doesn’t need precision. 65V on the collector, 15 on the base, and a 4.7 ohm resistor (4 in series parallel you need a 50 watter) in the emitter. The test does NOT need to be pulsed at all - just at a controlled temperature. A finned heat sink with part of the fins in an ice bath works. That will keep Tc at 25C or less. Scale the emitter resistor for smaller transistors. Use a bench supply for the base. A 48 volt transformer and rectifier is fine for the collector. Doesn’t even need to be filtered - just let the peaks hit 65-68 volts. Fuse (or current limit) the base at 1A, and fuse the collector at 7. If it blows it blows. Then you knows.
Screening vceo is easy. I start with a 340VDC power supply and a 100k resistor. Simply measure the voltage you get, and stay well under. Can sort unmarked zeners and unknown transistors (or ones the marking has worn off) that way too so it’s a good idea to leave that rig together too.
The problem with smaller transistors (ie 2SD699, 2SC3503) is that fakes will normally PASS the manufacturer’s published SOA. Look at their data sheets, the power handling isn’t wonderful. If you need to screen these types you need a way to test speed because thats where the differences are going to show up.
Not universal at all, but for my bread and butter Guitar amps I was forced to build a test set for 2N3055 class transistors, my basic building block, which started failing scandalously between 2004 and 2008, by which point they were unusable.
I tested for dataheet specs and "a little more"
Fed them 40Vce from a simple fixed supply, made them pass 2A, (2V across a 1 ohm emitter resistor), had a pushbutton to turn them ON.
Any which stood 15 seconds of it went to the "confirmed" box.
To test many quickly, I mounted them to a flat heatsink (a floor tile sized 3mm thick square aluminum plate) with a quick clamp, dropped a drop of oil as "heat sink grease" instead of messy "bird poop" and had a small plug-in connector for BE terminals.
As time went by, the reject pile grew and grew ... not funny.
Guess it can be scaled for larger power transistors, it was good enough for me back in the day.
The fake conspiracy's grow bigger
But relentless drive into 4 ohms is fun.
Pile on as many output devices as possible.
Buy OnSemi from usual distributors.
Digi or Mouser.
Seems like people want things to be fake.
To solve fake issues.
More a issue depending on region and distributors.
If USA location, rather easy to obtain authentic devices
But relentless drive into 4 ohms is fun.
Pile on as many output devices as possible.
Buy OnSemi from usual distributors.
Digi or Mouser.
Seems like people want things to be fake.
To solve fake issues.
More a issue depending on region and distributors.
If USA location, rather easy to obtain authentic devices
It WAS easy to obtain authentic Onsemi devices - about 5 years ago. Don’t know where the industry is headed now. Today you need plan out projects and have two year (or more) time frames for completion. Parts will come back into stock eventually. What happens 5 years from now? Will EVERYTHING go MOQ with 6-18 months lead time? I just ordered some Sankens that are for a project which might get completed a year or two from now, because the original type had gone obsolete already, and the replacement was down to a handful at Digikey - and they weren’t taking back orders. Others had already gone MOQ - I figured these darlingtons were next on the chopping block. There is no goddam room for a separate driver, PCBs are already here, and the TIP142/7 don’t have enough SOA. And if I didn’t get them right then I’d need to either redesign or order 250 of them at some point. You look in Mouser right now, and the only two pairs in stock are TIP35/6A or MJL4281/4302. The rest are out of either the NPN or PNP. You can mix-match packages, but you aren’t getting the same till April. Some longer than that.
Not much point if the SOA is really low on the fake device it will blow anyway.
I tested my amp at 1 watt 2 83v audio signal loaded with 0.22 ohms switched in and out.
No audible sound quality change and no blown outputs.
I tested my amp at 1 watt 2 83v audio signal loaded with 0.22 ohms switched in and out.
No audible sound quality change and no blown outputs.
And for the most part
On Semi still sells TO-264 packages
When more common smaller TO-247 and TO-3P-3
being the usual.
Hilarious they have 230 watt power transistors
which can average 35 MHz Ft and go up to 64 MHz
min gain is 80 not the max. Power transistors
way up to 250 hFE
No better time to be alive and have fun with BJT.
You do realize they manufacture here in USA
33 Acre construction project.
There is no fake BS
Military, Aerospace Manufacture dont make fakes LOL
Product comes with a Datasheet
Which includes manufacture location codes
And traceability code stamped on the device
3 leads , solder and done
On Semi still sells TO-264 packages
When more common smaller TO-247 and TO-3P-3
being the usual.
Hilarious they have 230 watt power transistors
which can average 35 MHz Ft and go up to 64 MHz
min gain is 80 not the max. Power transistors
way up to 250 hFE
No better time to be alive and have fun with BJT.
You do realize they manufacture here in USA
33 Acre construction project.
There is no fake BS
Military, Aerospace Manufacture dont make fakes LOL
Product comes with a Datasheet
Which includes manufacture location codes
And traceability code stamped on the device
3 leads , solder and done
- Home
- Amplifiers
- Solid State
- Output transistor load testing?