Oscilloscopes are Exciting

[xoaphexox]

So I have been messing around with restoring vintage audio for almost 3 years now but had never gotten in to anything but recapping, cosmetic restoration and simple repairs.

I decided it was time to get to the next level and with the help of our own Dr*Audio I was able to get an oscilloscope and signal generator. I recently turned half my basement in to a workshop area and I feel the bench is now complete!

This weekend I decided to start my new process of measuring the power output of my amps, out of curiosity and to learn how to use the oscilloscope.

I decided to start with a Sansui AU-999 that I recently completely restored - all new capacitors, fuses, and I went through the painstaking process of replacing every single "flying saucer" capacitor with new ones.

I am not sure if I am doing things exactly correct so here is where I will describe what I did and I would appreciate if anyone could chime in and make suggestions.

1. I hooked up a 200W 8 ohm dummy load to the right channel of the AU-999 and verified its resistance at 8.0 ohms.
2. I made a home-brew RCA cable to go from the signal generator to the right channel input of the AU-999.
3. I turned the signal generator up as high as it would go (1.65V) and set it to 1kHz
4. I turned on the scope set to 10V/division and AU-999 and let them warm up for several minutes.
5. Then I slowly turned up the AU-999 while observing the sine wave until it went in to clipping. I then lowered the volume to see it look good again and repeated this about 3 times.
6. I recorded the maximum voltage of the channel at 38V.


So by this process P=E^2/R:

P = 38^2/8 = 180.5W

To me this figure seems very high considering the AU-999 rates itself as a 50W/50W continuous power output at 8 ohms and 140W "IHF Music Power" at the same resistance.

Now I realize I was only driving one channel - is there a way to figure out what both channels would be or should I measure them simultaneously?

I am trying to determine the best way to hook the signal generator up to both channels simultaneously - I might have to hack together a wire for that - what do you guys do?

It has been a fun learning experience and I look forward to getting the output measurements of all the units in my basement!

Peace! :music:

https://home.comcast.net/~devanymarino/bench.jpg

[Hyperion]

Did you measure the output voltage with your new toy, i.e. the scope or with a true RMS ac voltmeter ? :scratch2:

The scope will give you the peak to peak voltage.
The voltmeter will give you the RMS voltage.

The difference in these values will alter your power calculation considerably, your formula is correct for the RMS voltage.

Just my 3 ha'pence worth

John

[jstang]

You need to convert to RMS Voltage by multiplying the PEAK voltage by .707 before doing the P=E^2/R.....

90 WPC RMS. Still seems a bit high, but I am not familiar with Sansui units.


Make sure the scope is calibrated....and that your are in deed measuring from the ground point on the scope display.


There calibration knobs on the voltage ranges usually are the smaller concentric knob.

jk

[xoaphexox]

Jstang and Hyperion: Thanks greatly for telling me about the RMS voltage difference.

I realize "both channels driven" will give a bit smaller reading so 90W out of one channel doesn't seem terribly off considering the shape of the unit and the tendency for Sansui to downplay their power measurements during that period. Not to mention this is just a 1kHz power rating and not the 20Hz to 20kHz that is mentioned in the service manual of being 50W.

Either way its a good starting point and I look forward to taking more measurements.

It is my understanding the scope is calibrated because Dr*Audio gave it a going-through before he sold it to me.

[xoaphexox]

By the way I must have measured wrong last time - I just measured again and was one less division than before, so with RMS I get 78.5W, that seems much more realistic!

Thanks for the tip!

[EchoWars]

Every 999 I've ever measured struggled to hit its rated 50W.

[xoaphexox]

Every 999 I've ever measured struggled to hit its rated 50W.

In this AU-999 I replaced:

37 electrolytic capacitors with Panasonic FC
16 Mylar capacitors with Panasonic Metal Film Polypropylene
2 Polypropylene capacitors
30 transistors
4 fuses

And if I am not mistaken it was under your guidance, Echo.

If it should struggle to reach 50WPC can you offer more advice or guidance as to what I might be doing wrong? I eventually intend to sell some of my amps as money is tight right now and I don't want to mislead anyone. I planned to use the power output rating as a selling point.

Thank you.

[EchoWars]

You need a real, True RMS-reading meter to be measuring things like maximum output power (or a distortion analyser). Plus, you need to be loading down both channels when doing measurements, not just one.

50W = 20VRMS into 8 ohms

[skippy_ps]

From looking at that pic, I assume you're not married.

Murray

[jstang]

Most scopes have an 1Vp-p 1KHz square wave reference signal output. Practice making different measurements on that signal. Also should help tell if the the scope is measuring correctly.

jk

[EchoWars]

Reading the voltages on a scope (unless it has a digital display to tell you the P-P voltage or RMS) leaves too much room for interpretation. You need a meter that reads true RMS, and can deal properly with a 1KHz signal. Too much work and too error-prone to read scope divisions and interpolate and calculate RMS voltage when a decent meter will do it for you.

[oldhifiguy]

that's the 1st time in all the yrs I've used one have I heard of a scope being called "exciting".

[Hyperion]

I think they are exciting too ! - so make that #2 oldhifiguy :D

My scope is a hybrid !!! - yes Tubes AND Transistors, it's nearly 40 years old now so needs some reconditioning work, however I don't have a service manual :tears:, so likely, if it developed a fault, it would become a door stop. It's dual beam and 5Mhz bandwidth - all I need now is a decent scope probe.

Oops sorry for the hijack...

John

[xoaphexox]

From looking at that pic, I assume you're not married.

Murray

Ha! Actually I'm engaged and will be getting married in about 2 weeks.

The bench is in the basement - when we bought this house we each chose a room as our own that the other was not allowed to decorate or make decoration suggestions about. She has her "craft room" upstairs and I have my "man cave" in the basement :music:

[xoaphexox]

You need a real, True RMS-reading meter to be measuring things like maximum output power (or a distortion analyser). Plus, you need to be loading down both channels when doing measurements, not just one.

50W = 20VRMS into 8 ohms

Good advice. So I need an RMS voltage meter to take voltage readings right before the wave starts to clip.

I do have another 200W 8 ohm dummy load that I will use to load the other channel next time I take a reading.

What do you usually use to split the signal generator output between the two channels? My Tenma signal generator only has one channel output. Is it safe to simply splice the output using alligator clips??

Thanks again.

[EchoWars]

Is it safe to simply splice the output using alligator clips??Why wouldn't it be? Unless the Tenma has some really piss-poor output impedence, you ought to be able to feed a lot more than two channels.

[xoaphexox]

Thanks for the input, Echo. This signal-feeding and output level-reading is new territory for me so I apologize if I don't quite understand the implications of each step of the process quite yet.

I must say though the yearly $25 subscription fee for AK is more cost effective than some college courses I have attended. I greatly appreciate your willingness to help me understand things.

It had been suggested that one could ballpark RMS by multiplying your voltage by .707 - what do you have to say about that?

[jstang]

http://ourworld.compuserve.com/homepages/g_knott/elect39.htm


.707 ---- It is not a ballpark suggestion...it is the calculation that is used to convert a pure Sine wave Peak measurement to RMS. .707 x Vp is the RMS (Root means squared ) calculation.

The problem is that SEEING the exact point of clipping is not always easy on a scope....and accurate... But who cares really, as long as you are close.

Distortion meter is the best method.

Myself, I have my doubts in using a simple true RMS meter. As distortions can be taking place and how is the meter notifying the user....

Clipping is distortion and the meter will not be able to notify you that it is taking place....

That is where the distortion meter comes in to play...to detect when the pure sine wave is being clipped....


jk

[EchoWars]

Sure...but not everyone wants to invest in a distortion analyser. Turning up the output till you just barely see the waveform beginning to clip and reading the RMS voltage on a meter is good 'nuff for Gov't work.

[xoaphexox]

Okay, I looked at RMS Flukes and they are in the $300 range so I have no choice but to use the poor man's method for now of using a $40 multimeter to read voltage in combination with reading the scope output and then multiplying (the average?) of those two readings (even though they should be the same) by .707 before squaring it and dividing by the resistance.

I will describe my method of measurement as the disclaimer to any power output statements that I post online, that should be a good enough C.Y.A. for me until I can afford a distortion analyzer or a RMS Fluke.

Thanks for your help guys!

:thmbsp:

[jblmar]

In your case, xoaphexox since you're not using a distortion analyzer, I would back off on the amp's output just a little more after you see clipping on your o'scope. Distortion rises very quickly at the onset of clipping. You may get a high RMS reading but the distortion level may be much higher than specified.

Ron

[EchoWars]

Okay, I looked at RMS Flukes and they are in the $300 rangehttp://cgi.ebay.com/Used-Fluke-179-True-RMS-multimeter_W0QQitemZ200233769734QQihZ010QQcategory Z58277QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
http://cgi.ebay.com/FLUKE-179-Digital-True-RMS-Multimeter-Excellent-Cond_W0QQitemZ230261685458QQihZ013QQcategoryZ58277 QQrdZ1QQssPageNameZWD1VQQcmdZViewItemQQ_trksidZp16 38Q2em118Q2el1247

[xoaphexox]

http://cgi.ebay.com/Used-Fluke-179-True-RMS-multimeter_W0QQitemZ200233769734QQihZ010QQcategory Z58277QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
http://cgi.ebay.com/FLUKE-179-Digital-True-RMS-Multimeter-Excellent-Cond_W0QQitemZ230261685458QQihZ013QQcategoryZ58277 QQrdZ1QQssPageNameZWD1VQQcmdZViewItemQQ_trksidZp16 38Q2em118Q2el1247

I would pass on the first one (62% feedback), but do you think used DMM's are a good idea? How easy would it be for one to go out of spec during its lifetime? Can they be recalibrated? I hesitate to spend even $160 on something that might still give me inaccurate readings.

Thanks for taking the time to try to help me out though, I do appreciate it.

[jstang]

Sure...but not everyone wants to invest in a distortion analyser. Turning up the output till you just barely see the waveform beginning to clip and reading the RMS voltage on a meter is good 'nuff for Gov't work.


I agree with that approach....that is how I do it.... I have a real scope and a hand held...

The hand held has a built in Meter...so I get two for one as they say.....


I picked up a distortion meter for 120$.....

All good enough for government work....:D

Well not really, some gov work is very very demanding....but good enough for my basement....


jk

[EchoWars]

I would pass on the first one (62% feedback), but do you think used DMM's are a good idea? How easy would it be for one to go out of spec during its lifetime? Can they be recalibrated? I hesitate to spend even $160 on something that might still give me inaccurate readings.I've carried Fluke meters banging around in my toolbox and severely mistreated by the airlines for 20+ years. Take a licking and keep on ticking. And they rarely go out of cal (but can be recalibrated if you suspect something wrong).

[d3imlay]

A non true RMS meter will probably read sine waves with reasonable accuracy, at least to 1Khz. Any non sinusoidal waveform will not be accurate. An HP 334 analyzer can often be had for around $100. There are tons of True RMS meters that come up on ebay. You can't go wrong with a Fluke. Pawn shops might be a good place also.

[othello]

How did you measure those 38 V you mentioned initially?
If this is a peak to peak value, then your calculation would have to be:

RMS= p-p/2 * 0.7

If you do the calculation with this value you get an output of around 20 watts.

I don't think you need additional gear, you just first have to establish exactly what you are measuring, afterwards the results you get from a scope are probably quite good enough for what you try to accomplish.

Uwe

[tensleep]

I am in the same mode that you are in - audio as a hobby. I studied electronics for a couple of years and have an EET Associates degree. However, my education led me into industrial maintenance and eventually to management. I want to brush up on my skills and even learn tube theory, so I have been building a test equipment inventory as well.


I am trying to determine the best way to hook the signal generator up to both channels simultaneously - I might have to hack together a wire for that - what do you guys do?

I intend to use my laptop, running WinISD, a free program, as my signal source. I can't remember where I downloaded the program, but you could probably search AK and find a website. The advantages of using your laptop as the audio test source are tremendous! Good luck and share your experiences, please!

[xoaphexox]

How did you measure those 38 V you mentioned initially?
If this is a peak to peak value, then your calculation would have to be:

RMS= p-p/2 * 0.7

If you do the calculation with this value you get an output of around 20 watts.

I don't think you need additional gear, you just first have to establish exactly what you are measuring, afterwards the results you get from a scope are probably quite good enough for what you try to accomplish.

Uwe

It is a peak voltage measurement as opposed to a Peak-to-peak.

I plan on taking more readings soon with a DMM in addition to the scope so I can see what's what.

[nashvillebil]

I picked up a new Fluke 179 from www.metersuperstore.com for a tick over $200. Good meter. (I got the electronics package, with leads etc. )

[fuzzy]

From looking at that pic, I assume you're not married.

Murray

LOL!!!!

I had a preamp + test gear on the kitchentable for 3 weeks and I finally got it finished yesterday. Looked much more busy then in this pic. I am married and my wife didn't say anything. In fact we ate lunch etc right next to my disassembled unit. Am I lucky or what :banana:

oh, and now I got some amp sitting on that kitchentable, oh no!

[fuzzy]

I think they are exciting too ! - so make that #2 oldhifiguy :D

My scope is a hybrid !!! - yes Tubes AND Transistors, it's nearly 40 years old now so needs some reconditioning work, however I don't have a service manual :tears:, so likely, if it developed a fault, it would become a door stop. It's dual beam and 5Mhz bandwidth - all I need now is a decent scope probe.

Oops sorry for the hijack...

John

Worked with one of those hybrids ovr 20 years ago and it was old then. Back then I thought a scope was "exciting". Today there are boring unless I am over at a friends house who collects scopes and has some really interesting pieces and knows how to use em too.

[d3imlay]

Come to think of it, I have a dual trace usb scope that I use for my electronics training classes. My primary application is to be able to demo real time waveforms via projector in the class room. Something cool about the scope is it also has a built in voltmeter for each channel that reads peak, RMS, average etc. Cost is about $200. Google "stingray scope"

[xoaphexox]

Tonight I went back to the bench and hooked up both dummy loads along with a DMM to each one. Here are the results:

28V * .707 = 19.796V ^ 2 = 391.881616 / 8 = 48.985202W

Looks like Echowars was spot on (not suprisingly)

I must admit, considering the time and money I invested in this unit I am shocked it doesnt live up to its rated 50W/CH even if its only by <1.02W. I realize I don't have laboratory quality equipment but its not junk either!

[nashvillebil]

Remember that as the dummy loads heat up, their resistance increases...so the figures may not be as accurate. According to a calculator I found on the Web, if the temp goes up from 50 to 100 degrees C, the resistance goes up to 9.5 ohms.

Also the dummy load may not be completely non-inductive...

[othello]

Tonight I went back to the bench and hooked up both dummy loads along with a DMM to each one. Here are the results:

28V * .707 = 19.796V ^ 2 = 391.881616 / 8 = 48.985202W


I love how absolutely precisely you were able to figure out the output, down to microwatts!!

Uwe

[EchoWars]

Remember that as the dummy loads heat up, their resistance increases...so the figures may not be as accurate. According to a calculator I found on the Web, if the temp goes up from 50 to 100 degrees C, the resistance goes up to 9.5 ohms.

Also the dummy load may not be completely non-inductive...Looks like he has the big 200W 8 ohm Parts Express dummy loads. I've had mine hot enough to take skin off your fingers, and the resistance did not change. Yes, they are non-inductive, and 50W into one of these would take a long time (30 minutes) to heat up the resistor.

[whoaru99]

Tonight I went back to the bench and hooked up both dummy loads along with a DMM to each one. Here are the results:

28V * .707 = 19.796V ^ 2 = 391.881616 / 8 = 48.985202W

Looks like Echowars was spot on (not suprisingly)

I must admit, considering the time and money I invested in this unit I am shocked it doesnt live up to its rated 50W/CH even if its only by <1.02W. I realize I don't have laboratory quality equipment but its not junk either!

I'd say that's pretty darn close - considering that you're not monitoring distortion to try to match exactly to the factory specification. Although, I'd have expected a bit more than the 20Hz-20kHz power rating when using only a 1kHz test.

[xoaphexox]

I'd have expected a bit more than the 20Hz-20kHz power rating when using only a 1kHz test.

Now that you mention it that is peculiar.

Out of curiosity, what main factors influence the WPC rating? Bias? I set it to spec a few months ago but I can re-check it.

[whoaru99]

Afaik, the main factors that determine WPC rating are the power supply rail voltages and the output stage configuration.

There's a lot of stuff between A and Z, but those would seem to be the predominant factors, IMHO.

[jstang]

I would double check your readings.

The way I do it on a scope, is let the signal clip a little measure the peak to peak.... Half that number and then (.707 x VP ) squared / load.....

Besides your DMM reading....

I think Echowars said he has never seen that model beat the rating....


jk

[xoaphexox]

Thanks for the advice Jstang, I will give that a whack once I get back from the 'shack with new alligator clips.

[EchoWars]

Amps nearly always do better than their listed power ratings...small amps by 10~20W, large amps by 100W or more. Thus I found it unusual to see this series of Sansui amps struggle to meet their rated wattage. By the standards set before and since then, the 999 ought to have been rated as a 35 or 40WPC amp.

My own method of testing for power output at clipping is to feed a 1KHz sine wave and push the output till I see 0.5% THD+N at the clipping point (1% is probably a more common figure, but I'm pickier). I then take an RMS reading of the voltage and calculate power. Using this method on the few 999's that have been here, I usually calculate right at about 50W.

[Susurus]

Sure 'scopes are fun! Fun to take apart.

Inside there is a triple-eye-stalked War Of The Worlds alien. :yikes:

Previous owner used that empty module bay for cinnamon buns? :dunno:

http://www.audiokarma.org/forums/attachment.php?attachmentid=98879&d=1215806677

Just took apart the Tek 7834 I lugged in last night (with 3 modules). Still couldn't get it to work. The switching PS tries to fire up; says "tick-tick-tick" but won't get anywhere...

bama.sbc.edu has the SM for it free, :thmbsp: but its server is a little leasurely. No matter I have other projects... :D

My 475 has the jitters from a cantankerous fan.

Meanwhile the Fluke 77 never fails. Ever. :music:

[Susurus]

SM says that little neon light is a warning that the PS caps are charged and >80V.

But why do the front indicator lights flash in concert with the tiny warning light? Arghh

[dshoaf]

I'd be looking at all the tantalum caps that decouple the voltage supply rails throughout the scope. All 3 of my HP 17xx series scopes have had this problem and its well known on the HP/Agilent Yahoo group, too.

I just went through and swapped them all out and lots of strange things stopped happening - and the power supply ran cooler, too. Also make sure all the AMP connectors are clean, too, as I had one of those on the power supply board melt down due to heat from bad contacts.

Cheers,

David

[Susurus]

Thanks I really appreciate it.

:thmbsp:
:thmbsp:
:thmbsp:

[xoaphexox]

Just got done putting my restored BA-90 under the scope. This one came out 11 wpc higher than stock using the previously mentioned method.

That's more like it!

https://home.comcast.net/~devanymarino/bascope.jpg