Thanks Charles, I found a similar problem about 20 years ago with the National Semi LF411, that I used, up till then, for Vendetta Research servos. I switched over to the AD711, and things improved considerably. I have a WHOLE plastic drawer of NOISY LF411's that I hoped to give away to somebody, someday, but SY has moved away ;-)
I now use the OPA134 in its place, I should actually check them, as I thought it was a National problem, almost exclusively.
I now use the OPA134 in its place, I should actually check them, as I thought it was a National problem, almost exclusively.
Now that discussion has dropped a little, I might point out something 'interesting' at least to me, that happened this last friday afternoon.
I had to do a new 'show and tell' for my 'boss' who came in from 'out of town' to see my progress on a new phono stage. I had the equipment all hooked up and running: An ST1700B and a HP3563 in series to get distortions below 110dB. The problem was my TEST LEADS. They are getting old and the clips on the end would NOT stay on the terminal they were attached to. After several disruptions, because of a slipping clip lead, my boss said that we should get some other test leads at the local electronics supply. So off we went, and he bought me some test leads, not the best test leads in the world, but the best that they had available, and we took them back to the lab. Well, I opened the package and got out a NEW test lead, and replaced the most troublesome one. It happened to be the test lead that supplied the test signal to the phono preamp gain stage. Mechanically, it was 'hog heaven' no more slips, BUT something else creeped in: HIGHER ORDER HARMONIC DISTORTION. At first we didn't see much, and I was operating the preamp 10dB about its maximum useful working level, but there was a little. Then my boss asked me to drop the input level to just peak output level. This is when the higher order distortion INCREASED! How can this be? A class A open loop differential stage does NOT increase with lower levels. Then he said to lower it further to operating level, and the distortion INCREASED even further! Now, I am in an embarrassing situation, where did the distortion come from? What to do? Well, I tried changing the connections and grounding of the system, so perhaps a miracle would happen, and it did! I found that when I REMOVED the external ground connection to the input of the ST1700B, the higher order distortion mysteriously disappeared. Well, that saved the day, except for WHY did the distortion happen in the first place? (more later)
I had to do a new 'show and tell' for my 'boss' who came in from 'out of town' to see my progress on a new phono stage. I had the equipment all hooked up and running: An ST1700B and a HP3563 in series to get distortions below 110dB. The problem was my TEST LEADS. They are getting old and the clips on the end would NOT stay on the terminal they were attached to. After several disruptions, because of a slipping clip lead, my boss said that we should get some other test leads at the local electronics supply. So off we went, and he bought me some test leads, not the best test leads in the world, but the best that they had available, and we took them back to the lab. Well, I opened the package and got out a NEW test lead, and replaced the most troublesome one. It happened to be the test lead that supplied the test signal to the phono preamp gain stage. Mechanically, it was 'hog heaven' no more slips, BUT something else creeped in: HIGHER ORDER HARMONIC DISTORTION. At first we didn't see much, and I was operating the preamp 10dB about its maximum useful working level, but there was a little. Then my boss asked me to drop the input level to just peak output level. This is when the higher order distortion INCREASED! How can this be? A class A open loop differential stage does NOT increase with lower levels. Then he said to lower it further to operating level, and the distortion INCREASED even further! Now, I am in an embarrassing situation, where did the distortion come from? What to do? Well, I tried changing the connections and grounding of the system, so perhaps a miracle would happen, and it did! I found that when I REMOVED the external ground connection to the input of the ST1700B, the higher order distortion mysteriously disappeared. Well, that saved the day, except for WHY did the distortion happen in the first place? (more later)
Well, I thought about the problem with the higher order distortion and thought that I might find out its 'cause' at least its source. So, I REMOVED the phono preamp from the circuit and ONLY put the input and output leads together. Voila! There was the higher order distortion, once again! So now it is the test gear or the test leads, or BOTH.
So, I removed the test leads and put in a 75 ohm cable that I knew measured well for this sort of problem, and the distortion DISAPPEARED with the same levels and loading.
This makes me suspicious that something in the NEW test lead is generating this problem. We shall see, in future.
So, I removed the test leads and put in a 75 ohm cable that I knew measured well for this sort of problem, and the distortion DISAPPEARED with the same levels and loading.
This makes me suspicious that something in the NEW test lead is generating this problem. We shall see, in future.
Were you monitoring with a scope or listening at the same time ?
You could have enough hum to drive the circuit out of its
linear region..........
Scott, thanks for jogging my memory. We have the older one with the bank of analog meters. They are much easier to read than digital meters.
Remember when the first cars came out with digital speedometers? That didn't last very long! (Don't those guys ever test their cars before they sell them? They would have realized instantly that a digital speedometer is a disaster...)
Test Lead distortion
Hi John,
I have seen this exact problem on my bench in the past. I was able to trace it to the crappy way the nickel plated steel alligator clip was crimped to the wire. There was no attempt to make a gas-tight crimp, the wire was merely folded over the insulation, and then the mess was crimped in the clip. Soldering the wire fixed the problem in my case. I suspect that chlorine from the PVC insulation had corroded the copper against the nickel, making a non-linear junction. Under a microscope, most copper wire with PVC insulation can be seen to have a greenish cast from this corrosion. I guess non-contaminating jackets are too expensive for cheap wire.
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill
www.wxyc.org
1st on the Internet
Hi John,
I have seen this exact problem on my bench in the past. I was able to trace it to the crappy way the nickel plated steel alligator clip was crimped to the wire. There was no attempt to make a gas-tight crimp, the wire was merely folded over the insulation, and then the mess was crimped in the clip. Soldering the wire fixed the problem in my case. I suspect that chlorine from the PVC insulation had corroded the copper against the nickel, making a non-linear junction. Under a microscope, most copper wire with PVC insulation can be seen to have a greenish cast from this corrosion. I guess non-contaminating jackets are too expensive for cheap wire.
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill
www.wxyc.org
1st on the Internet
Last edited:
Hhoyt, I think you have it! Except that it is the OLDER interconnect that appears to be the problem, not the new one. What a shock. Pomona, of all things, BUT an OLDER Pomona. It appears to be in the ground return that is the shield for the interconnect, as well. Copper leads with this interconnect. Cleaning and DeOxiting helps only a little.
Quan-Tech transistor Noise Analyzer
The older ones are pretty good for a mid-1960's design. They were well built which is why the are still used. Once they get flakey however the age of the components shows.
The digital readout version I have has different time constants for each meter so they are all quite readable. Its also a lot smaller with a separate test head. And it can test a quad op amp from buttons at the display. However you need to manually move the transistor in the socket to measure a dual transistor.
It has an output that can be connected to a spectrum analyzer.
The older one needs a different control unit for IC's (actually a different product).
Having used both I prefer the newer design (mid 1980's). Mine is now 23 years old. Time for a new generation.
The older ones are pretty good for a mid-1960's design. They were well built which is why the are still used. Once they get flakey however the age of the components shows.
The digital readout version I have has different time constants for each meter so they are all quite readable. Its also a lot smaller with a separate test head. And it can test a quad op amp from buttons at the display. However you need to manually move the transistor in the socket to measure a dual transistor.
It has an output that can be connected to a spectrum analyzer.
The older one needs a different control unit for IC's (actually a different product).
Having used both I prefer the newer design (mid 1980's). Mine is now 23 years old. Time for a new generation.
Analog: model 2283 (op amp noise analyzer control unit) or 2173-C (transistor noise analyzer control unit) used with model 2181 (noise filter unit)
Digital: model 5173
QuanTechs
See, I told you Demian was a test equipment guru!
And we also have another QuanTech guru.
If only someone would build a new product that did the same thing for less than $40,000. I wanted to buy a phase noise analyzer for digital work but had to blink a few times when I saw the sticker price -- $92,000.
Of course it is tax-deductible! But I don't think I would use it enough to justify it. Probably better off to rent it for a month for a few thousand dollars to see if it is really something I would need.
See, I told you Demian was a test equipment guru!
And we also have another QuanTech guru.
If only someone would build a new product that did the same thing for less than $40,000. I wanted to buy a phase noise analyzer for digital work but had to blink a few times when I saw the sticker price -- $92,000.
Of course it is tax-deductible! But I don't think I would use it enough to justify it. Probably better off to rent it for a month for a few thousand dollars to see if it is really something I would need.
For audio work you don't need the $92K system. And even with that you still need a SOTA oscillator at the target frequency.
You can build an SOTA phase noise system with two oscillators (the one you want to test and a reference oscillator at the target frequency), a double balanced mixer, a low noise amplifier and a spectrum analyzer. There are a few other bits but its not that expensive if you are not trying to qualify microwave sources. This link Techniques for Measuring Phase Noise will guide you to getting results the same way NIST does. And the low noise amp they propose could be replaced with an AD797. The bigger challenge is getting an ultra low noise reference oscillator. Wenzel will provide for about $1500 per frequency, still a lot less than $97K. Or this is cheaper and a newer design for a packaged system: Holzworth - Analyzers
Not much chance of that. One of our QC managers visited them (him!) to try and get ours updated and calibrated. He offered us the whole business because none of his kids wanted it. It makes no sense these days, an NI card and Labview could measure/analyze the noise in any way you wanted.
- Status
- Not open for further replies.