Hello everyone, new here! Happy to be here, nice to meet everyone. First off, very entry level DIY here. Engineering background but that really does not mean much. I can read schematics, use tools and do math so I suppose that counts for something.
In this process I am trying to learn as I go. Here is the issue.
The other day my amp started buzzing, in denial I turned it on and off a few times (because I bought it under the pretense that it was just serviced) and it ended up blowing my woofer and the main fuse in the amp. From some rudimentary googling I somehow got to the point of testing output transistors. I have been studying the schematic for over an hour now and I see there are 2 output transistors per channel, two driver transistors and then maybe predriver transistors (still learning). Anyhow, I know the driver and output transistors are NPN. First question is, can i test in place?
If so here are the readings but what bothers be is why are ALL of the outputs shorting, something else must have caused this? Before just replacing them I wanted to see if anyone has any advice. Thank you in advance for your help!
Outputs:
Positive lead to Base, Negative to Emitter (ALL ARE SHORTED)
Positive lead to Base, Negative to Collector (voltage drop of about 0.425)
Drivers:
Positive lead to Base, Negative to Emitter 0.15V drop
Positive lead to Base, Negative to Collector 0.450V drop
In this process I am trying to learn as I go. Here is the issue.
The other day my amp started buzzing, in denial I turned it on and off a few times (because I bought it under the pretense that it was just serviced) and it ended up blowing my woofer and the main fuse in the amp. From some rudimentary googling I somehow got to the point of testing output transistors. I have been studying the schematic for over an hour now and I see there are 2 output transistors per channel, two driver transistors and then maybe predriver transistors (still learning). Anyhow, I know the driver and output transistors are NPN. First question is, can i test in place?
If so here are the readings but what bothers be is why are ALL of the outputs shorting, something else must have caused this? Before just replacing them I wanted to see if anyone has any advice. Thank you in advance for your help!
Outputs:
Positive lead to Base, Negative to Emitter (ALL ARE SHORTED)
Positive lead to Base, Negative to Collector (voltage drop of about 0.425)
Drivers:
Positive lead to Base, Negative to Emitter 0.15V drop
Positive lead to Base, Negative to Collector 0.450V drop
Hi solderbrain,
Well, what I would do in service is replace all the outputs and drivers. If one output goes, the opposite ones are stressed beyond their ratings. Also, the mate to the one that shorted may be on the way out and it was just a race to see who died first. Semiconductors can just fail for no operational reason as well. If it was a driver that died first, it was simply a case of "Who died first?".
Generally speaking you want to test the transistor out of circuit. You do want to measure beta and leakage current from C-E and C-B. THe cheap testers do not measure low enough currents. The best instrument I have seen so far is the Heathkit IT-18. It's fast and tells you exactly what you need to know.
Do not mix output transistor types or families. Continue checking upstream until you start hitting good stuff. I haven't looked at this schematic (I have repaired many), but my general advice to a new tech is replace the first good part you see as well (on your way back). Now also, the diff pair should be replaced because often one of the two will have gone into E-B reversed bias. That changes the characteristics often, and can make one noisy on top of that. They are supposed to be matched in beta.
Measure the resistors in the output section. Don't rely on visual inspection. They can open without showing any burning at times. Never assume anything, prove any assumptions you may have. Also beware of intermittent bias controls (trimmer), check them noting where they were set to first.
Well, what I would do in service is replace all the outputs and drivers. If one output goes, the opposite ones are stressed beyond their ratings. Also, the mate to the one that shorted may be on the way out and it was just a race to see who died first. Semiconductors can just fail for no operational reason as well. If it was a driver that died first, it was simply a case of "Who died first?".
Generally speaking you want to test the transistor out of circuit. You do want to measure beta and leakage current from C-E and C-B. THe cheap testers do not measure low enough currents. The best instrument I have seen so far is the Heathkit IT-18. It's fast and tells you exactly what you need to know.
Do not mix output transistor types or families. Continue checking upstream until you start hitting good stuff. I haven't looked at this schematic (I have repaired many), but my general advice to a new tech is replace the first good part you see as well (on your way back). Now also, the diff pair should be replaced because often one of the two will have gone into E-B reversed bias. That changes the characteristics often, and can make one noisy on top of that. They are supposed to be matched in beta.
Measure the resistors in the output section. Don't rely on visual inspection. They can open without showing any burning at times. Never assume anything, prove any assumptions you may have. Also beware of intermittent bias controls (trimmer), check them noting where they were set to first.
Anatech!
Thank you so much for this great information! Okay all well noted, I have to do some more research and learning on the beta and leakage, but really before you mentioned them, I was not even familiar with this.
Do you have any recommendations on where to purchase components? Ebay?
Also, worth noting, the reason I am trying to test in place is because all of the transistors are soldered in place, not sure if some can have socket receptacles but just figured to mention.
Thank you again so much for your time and input!
Thank you so much for this great information! Okay all well noted, I have to do some more research and learning on the beta and leakage, but really before you mentioned them, I was not even familiar with this.
Do you have any recommendations on where to purchase components? Ebay?
Also, worth noting, the reason I am trying to test in place is because all of the transistors are soldered in place, not sure if some can have socket receptacles but just figured to mention.
Thank you again so much for your time and input!
Never, ever, turn electronics on and off and on and off like that. It's very bad for the circuitry.
Hi solderbrain,
Rayma is correct, especially power switches and main filter caps (never mind the poor rectifiers!).
Only ever buy replacement parts from authorized distributors like Digikey and Mouser (or similar). Do not buy ECG or NTE replacement parts. You can find a modern substitution for those old devices. You may have to re-compensate the amplifier (feedback cap) for stability. You'll need an oscilloscope (analogue please) to check for oscillation or ringing. Input a square wave (about a volt output) should do it.
Ebay is chock full of fakes and re-marked parts sold for prices far higher than the part is worth most often. Today's transistors are better in every way compared to the originals.
Rayma is correct, especially power switches and main filter caps (never mind the poor rectifiers!).
Only ever buy replacement parts from authorized distributors like Digikey and Mouser (or similar). Do not buy ECG or NTE replacement parts. You can find a modern substitution for those old devices. You may have to re-compensate the amplifier (feedback cap) for stability. You'll need an oscilloscope (analogue please) to check for oscillation or ringing. Input a square wave (about a volt output) should do it.
Ebay is chock full of fakes and re-marked parts sold for prices far higher than the part is worth most often. Today's transistors are better in every way compared to the originals.
Great advice again, thank you.
Can you please recommend a good midlevel Oscilloscope, brand and retailer to purchase from? I remember using really nice ones in college and was always intrigued to learn to use them again! (College was 20 years ago)
Surprisingly the Crown service manual has great instructions on how to perform certain tasks so I am studying that tonight but will also spend some time reading and watching on the points you made above.
Can you please recommend a good midlevel Oscilloscope, brand and retailer to purchase from? I remember using really nice ones in college and was always intrigued to learn to use them again! (College was 20 years ago)
Surprisingly the Crown service manual has great instructions on how to perform certain tasks so I am studying that tonight but will also spend some time reading and watching on the points you made above.
Great!
Buy a used 'scope. 100 MHz, dual channel or more. You do want analogue! Philips had great traces, any Leader, B&K or similar (Hitachi are very good). Stuff like that. Look for burned screens or fuzzy traces. Some Teks are good, but I have a 2465B I do not care for at all. A 2235 or similar would be very good. I have a Philips PM 3070 and PM 3365A I love, and some older HP's that are great (I like them better than the Tektronix products).
In order to get trace quality with analogue signals, I bought a Keysight MSOX3104T loaded. My analogue scopes show an eye pattern much better, but this one captures and plays back waveforms and a bunch of other neat tricks. What drove me to this scope was the ability to show good eye patterns, it also has far less "blind time" for acquisitions.
DSO's have some tricks they do, but they pack a lot of "gottcha's" too. The cheaper ones can alias (I guess any can) and you must be aware of how they operate to get a true picture, or they will lie to you. Same for cheap meters. I have a T&M background as well. I also found the Keysight (Agilent / HP) product more intuitive to operate. You can't fight your gear and troubleshoot!
Buy a used 'scope. 100 MHz, dual channel or more. You do want analogue! Philips had great traces, any Leader, B&K or similar (Hitachi are very good). Stuff like that. Look for burned screens or fuzzy traces. Some Teks are good, but I have a 2465B I do not care for at all. A 2235 or similar would be very good. I have a Philips PM 3070 and PM 3365A I love, and some older HP's that are great (I like them better than the Tektronix products).
In order to get trace quality with analogue signals, I bought a Keysight MSOX3104T loaded. My analogue scopes show an eye pattern much better, but this one captures and plays back waveforms and a bunch of other neat tricks. What drove me to this scope was the ability to show good eye patterns, it also has far less "blind time" for acquisitions.
DSO's have some tricks they do, but they pack a lot of "gottcha's" too. The cheaper ones can alias (I guess any can) and you must be aware of how they operate to get a true picture, or they will lie to you. Same for cheap meters. I have a T&M background as well. I also found the Keysight (Agilent / HP) product more intuitive to operate. You can't fight your gear and troubleshoot!
Okay so first thing, the MSOX3104T, 22000 dollars?! Wow, I am not sure I can get to that level, ever! Haha 🙂
I am assuming from a quick search the other ones you mentioned should be able to be found for under 1000?
Can you, or anyone recommend a reputable reseller of scopes?
I found this place:
PM3070 Philips Analog Oscilloscope - Philips - Manufacturers (valuetronics.com)
The search begins! Thank you again!
I am assuming from a quick search the other ones you mentioned should be able to be found for under 1000?
Can you, or anyone recommend a reputable reseller of scopes?
I found this place:
PM3070 Philips Analog Oscilloscope - Philips - Manufacturers (valuetronics.com)
The search begins! Thank you again!
Last edited:
That's the one, with all software and options loaded. In Canada the price is higher. I really did not want to spend that much, but it turned three instruments into one on the bench and added capability. This is a budget 'scope. You don't want to know what else is on my bench.
Absolutely less than $1K, probably a couple hundred. The scope you linked to has the old colours like the one I bought, later ones have a lighter sand / tan type colour face. Mine was $2,500 new. HP had gone digital then and those wouldn't cut it. They were the best digital scopes, but not for analogue use.
Newer colours : https://www.ebay.ca/itm/14483957861...KZuIZvAsPf2THqOsJ3WTeM/g==|tkp:Bk9SR_K98aD_Yg
The PM 3365A looks almost the same : https://www.ebay.ca/itm/11570413400...ECpiSOzEYGqpr5635AahxnUQ==|tkp:Bk9SR9iQ-6D_Yg
If you can find one of these ... Look under rocks and with surplus dealers. Some of the "combiscopes" had a DSP chip that ran hot as heck. I'm not sure how they last. I would put a glue-on heat sink on them! I was offered one as a warranty replacement (I was having trouble with my new one) and I refused due to the temperatures inside.
So it looks like a premium 'scope should run you less than $500. Make no mistake, these are very nice oscilloscopes. Even an older model would be great.
Absolutely less than $1K, probably a couple hundred. The scope you linked to has the old colours like the one I bought, later ones have a lighter sand / tan type colour face. Mine was $2,500 new. HP had gone digital then and those wouldn't cut it. They were the best digital scopes, but not for analogue use.
Newer colours : https://www.ebay.ca/itm/14483957861...KZuIZvAsPf2THqOsJ3WTeM/g==|tkp:Bk9SR_K98aD_Yg
The PM 3365A looks almost the same : https://www.ebay.ca/itm/11570413400...ECpiSOzEYGqpr5635AahxnUQ==|tkp:Bk9SR9iQ-6D_Yg
If you can find one of these ... Look under rocks and with surplus dealers. Some of the "combiscopes" had a DSP chip that ran hot as heck. I'm not sure how they last. I would put a glue-on heat sink on them! I was offered one as a warranty replacement (I was having trouble with my new one) and I refused due to the temperatures inside.
So it looks like a premium 'scope should run you less than $500. Make no mistake, these are very nice oscilloscopes. Even an older model would be great.
Haha I can only imagine your bench 🙂
Okay great information again, thank you so much!
I have one more follow up question because I was searching on it and was not able to find any information:
You mentioned, "You may have to re-compensate the amplifier (feedback cap) for stability." can you kindly give me some direction where I may find more reading material on this topic?
Thank you again!
Okay great information again, thank you so much!
I have one more follow up question because I was searching on it and was not able to find any information:
You mentioned, "You may have to re-compensate the amplifier (feedback cap) for stability." can you kindly give me some direction where I may find more reading material on this topic?
Thank you again!
Amplifiers can oscillate all by themselves or on some loads on their speaker terminals. So often there are small capacitors across the feedback resistor or some other parts of the circuit. The value varies depending on phase shift through the amplifier and something called Phase Margin. You can study it, but the basic test is to run a square wave and see that it does ring on the leading (most common) or trailing edges. You also don't want it over-compensated where the edges of the square wave are rounded. Then you place loads across the output terminals to ensure it remains stable. Adjust the capacitance to correct the response. I use a trimmer capacitor and parallel with aa fixed one to get me in the range. Then I measure the required capacitance.
Also, allow the amplifier to warm up as that can change the stability as some parts warm up. The balance is good performance warmed up since it will most often be at running temperature, but not oscillating at any point. I can talk about Bode plots, but this is faster and effective.
Also, allow the amplifier to warm up as that can change the stability as some parts warm up. The balance is good performance warmed up since it will most often be at running temperature, but not oscillating at any point. I can talk about Bode plots, but this is faster and effective.
Excellent! This should be enough to keep me busy for the long weekend, thank you sincerely for all of your time!
I just glanced at the schematic. C205 is the main compensation cap. However (like many amps back then) there are more that will affect stability.
Also, replace the doubler capacitors. That's C4 and C5. Use axial leaded capacitors for the original axial types, do not bend the leads on a radial cap 'cause they are less expensive to buy. Also, if the newer capacitor bodies are smaller, increase the voltage rating to get the right sized case. I do not trust any capacitor rated under 16 VDC, and the higher the breakdown voltage is, the better quality capacitor you normally have. 63 VDC to 100 VDC is the sweet spot, but a 35 V cap will be vastly better than a 16 V cap. The most important thing is that the capacitor fits properly!
Modern capacitors have come a very long way. They are so much more reliable! That's assuming current stock, good brand etc ... No, coupling capacitors do not sound different to each other. I would replace all bypass capacitors, they are important and can cause an amplifier to oscillate if they are degraded.
Also, replace the doubler capacitors. That's C4 and C5. Use axial leaded capacitors for the original axial types, do not bend the leads on a radial cap 'cause they are less expensive to buy. Also, if the newer capacitor bodies are smaller, increase the voltage rating to get the right sized case. I do not trust any capacitor rated under 16 VDC, and the higher the breakdown voltage is, the better quality capacitor you normally have. 63 VDC to 100 VDC is the sweet spot, but a 35 V cap will be vastly better than a 16 V cap. The most important thing is that the capacitor fits properly!
Modern capacitors have come a very long way. They are so much more reliable! That's assuming current stock, good brand etc ... No, coupling capacitors do not sound different to each other. I would replace all bypass capacitors, they are important and can cause an amplifier to oscillate if they are degraded.
OK, I'll be the outlier on the scope. If you are only going to buy one, and you want a scope for any digital work, I'd go digital. I've got a inexpensive 2 channel rigol (1000 series) and now a 4000 series. The ability to trigger on I2C, SPI etc is essential to do any sort of digital. An analog scope needs the signal to be repetitive. Samplers grab it and display. I get sampler's have their own issues like not seeing stuff because of undersampling and the 8 bit quantizing. But for audio, a 1Giga Sampler is going to catch amp oscillations, and spurious ones too. The 8 bit thing is the hardest to deal with, one of the main reasons I went up to the 4000. That and I wanted 4 channels.
THANK YOU SO MUCH! Okay just a couple things I need to ask you to please clarify:
So, you are saying, instead of buying a radial cap and bending the leads, just buy a radial leaded caps?
If the new ones I buy to replace the older ones, dimensionally are smaller, up the voltage (which in turn will increase the size so that it matches)?
Is there an easy way to determine which are bypass capacitors? Or do you know know by where they are in the circuit? Are these also known as de-coupling?
So, you are saying, instead of buying a radial cap and bending the leads, just buy a radial leaded caps?
If the new ones I buy to replace the older ones, dimensionally are smaller, up the voltage (which in turn will increase the size so that it matches)?
Is there an easy way to determine which are bypass capacitors? Or do you know know by where they are in the circuit? Are these also known as de-coupling?
Also, I realize I am asking a lot of dumb questions, dumb as in basics probably in the world of amp design and circuitry. Can you recommend a resource for basics/101, either way I will work on just some good youtube videos, but of course if you or anyone have suggestions happy to have some varying sources 🙂
Hi Mike,
Having gone through the DSO range, the cheap ones are useless for audio. I have an Agilent 54642D, better than the cheap digitals today. It did okay on audio. It was a 500 MHz DSO. The cheaper DSOs I have are - useless.
If you want to do digital work and capture events, many will do. Keep in mind that catching a glitch will cost you for a scope capable of doing it in reasonable time. Many digital scopes are blind much of the time. I caught more with an analogue scope. Until I got the Keysight, I had two scopes on the bench, and I used the analogue one the most. The Keysight is 12 bit in high res mode.
Basically, if you want to do much in analogue with a digital scope, you're going to pay huge! There is also the matter of noise, the cheaper products are far noisier than an analogue by far. That really matters.
DSO's are great for toys, capturing events and doing serial decoding (mine does also). But, to look at real signals, you can't beat an analogue for the money. I'd be much happier with a 4000 series or higher, but that isn't in the cards. I still use an analogue occasionally.
You can get cheap logic adapters that hang on your USB buss for decoding serial or digital lines. I mean cheap. So for doing digital, be more imaginative. You can't do both without dropping a ton of money to get a DSO that will do audio well. It's cheaper to get an analogue and digital, the two are less expensive by far! For me, my bench has 43 instruments on it, and others I don't have room for that come out when needed. Plus two soldering irons and a pair of speakers and a GPS receiver in behind. Two DACs in behind also. There are probably other things there I didn't count. So for me, space is at a premium. I didn't count the computer and two monitors. FM antenna amp too. Then there are the various probes (current clamp types) and hand held meters. By comparison you fellas do have the room for both.
Having gone through the DSO range, the cheap ones are useless for audio. I have an Agilent 54642D, better than the cheap digitals today. It did okay on audio. It was a 500 MHz DSO. The cheaper DSOs I have are - useless.
If you want to do digital work and capture events, many will do. Keep in mind that catching a glitch will cost you for a scope capable of doing it in reasonable time. Many digital scopes are blind much of the time. I caught more with an analogue scope. Until I got the Keysight, I had two scopes on the bench, and I used the analogue one the most. The Keysight is 12 bit in high res mode.
Basically, if you want to do much in analogue with a digital scope, you're going to pay huge! There is also the matter of noise, the cheaper products are far noisier than an analogue by far. That really matters.
DSO's are great for toys, capturing events and doing serial decoding (mine does also). But, to look at real signals, you can't beat an analogue for the money. I'd be much happier with a 4000 series or higher, but that isn't in the cards. I still use an analogue occasionally.
You can get cheap logic adapters that hang on your USB buss for decoding serial or digital lines. I mean cheap. So for doing digital, be more imaginative. You can't do both without dropping a ton of money to get a DSO that will do audio well. It's cheaper to get an analogue and digital, the two are less expensive by far! For me, my bench has 43 instruments on it, and others I don't have room for that come out when needed. Plus two soldering irons and a pair of speakers and a GPS receiver in behind. Two DACs in behind also. There are probably other things there I didn't count. So for me, space is at a premium. I didn't count the computer and two monitors. FM antenna amp too. Then there are the various probes (current clamp types) and hand held meters. By comparison you fellas do have the room for both.
Hi solderbrain,
No. Replace like with like. So replace radial leaded with the same, and axial leaded with axial. I have seen people use radial leaded caps everywhere - no!!!!
Yes, exactly.
Yes, and yes. Some experience helps here. If in doubt, ask.