IMHO the diy'er has to be able to fix stuff. One of my heroes is Jim Williams of Linear Technology. He got so good by fixing and studying test equipment at MIT. Whatever you buy, it's going to need service and maintenance sooner or later. Given the age of affordable used test equipment, it will be sooner, probably the minute you get it.
Just getting things back to factory spec is a good start. As for calibration, you don't have to meet ISO9000, so just buy a few close tolerance voltage references and a few close tolerance caps and resistors. You can build up useful standards just from that. Worried about your frequency counter? Build a simple TRF receiver for WWVB at 60kHz. Frankly any scope that gets me within 1% is a very good scope indeed. As to the reference to the 465 scope, I can't remember how old they are, but by now the controls will be getting stiff and the failure rate will be rising. It was a good scope in its day, but it's heavy and is really ten pounds of you-know-what in a five pound sack. Not fun to work on. Give me a bigger boat anchor any day unless I have to carry it.
For teaching purposes I think an ADC like the USB modules from Measurement Computing would be far more useful. Prepare students for the future, not the past. You can set up a lot more interesting experiments when you can control something, automatically graph the results and ponder them.
Finally, you can probably improve the old THD analyzers, but they're already near the bottom of their lowest range. If you made a serious improvement, you'd need to add another range, or just use the output. My guess is improving the noise floor and improving the depth of null obtainable by enough to bother with, will be far more difficult than a diy'er can manage. You'd need better test equipment ;-)
Just getting things back to factory spec is a good start. As for calibration, you don't have to meet ISO9000, so just buy a few close tolerance voltage references and a few close tolerance caps and resistors. You can build up useful standards just from that. Worried about your frequency counter? Build a simple TRF receiver for WWVB at 60kHz. Frankly any scope that gets me within 1% is a very good scope indeed. As to the reference to the 465 scope, I can't remember how old they are, but by now the controls will be getting stiff and the failure rate will be rising. It was a good scope in its day, but it's heavy and is really ten pounds of you-know-what in a five pound sack. Not fun to work on. Give me a bigger boat anchor any day unless I have to carry it.
For teaching purposes I think an ADC like the USB modules from Measurement Computing would be far more useful. Prepare students for the future, not the past. You can set up a lot more interesting experiments when you can control something, automatically graph the results and ponder them.
Finally, you can probably improve the old THD analyzers, but they're already near the bottom of their lowest range. If you made a serious improvement, you'd need to add another range, or just use the output. My guess is improving the noise floor and improving the depth of null obtainable by enough to bother with, will be far more difficult than a diy'er can manage. You'd need better test equipment ;-)
Hi Conrad,
I completely agree, although I'm a believer that at least one instrument should be known to be in tolerance and the readings should be transferable to other instruments. People who are skilled at calibration can do more with less. So it is the beginner that really needs to confirm the instruments are telling the truth.
Hi bsgd,
I do have your email (just checked). I am having a rough time at the moment, but I do intend to respond once I'm feeling up to it. The HP 339A is better than some people may give it credit for. You made an excellent purchase. Did you down load the service manual for it yet?
Hi Joshua,
In my opinion, Ralph Nader was a headline seeking so and so. There were plenty of FORD products busy taking lives due to cost accounting measures, and yet he never spoke up once about that. Methinks he was under an agreement with FORD (a four letter word). I owned a couple F products that were actually dangerous, the Pinto was the best known. But then, all those cars where the transmission slipped out of park to kill people. Then there were all the gas line clamps that sprayed fuel and started fires.
My parents owned a pair of Corvairs (one was a Monza) that were excellent cars. The problem was that the weight was in the rear and people spun out do to a lack of skill coupled with poor driving decisions. Bad you say? Talk to me about a Porsche or VW bug.
The cars Ralphy boy killed were safer than the ones he allowed to take lives. I don't know how well known it is, but GM did more work in crash studies than anyone in North America. They implemented safety measures consistently before anyone else did on this rock. My Pinto truly was a poorly built POS that liked to drop it's gas tank on the road when it got hit from the rear (mine was). Luckily, the car was both out of gas and stolen when it got nailed.
You should study these "heros" before accepting what they apparently stood for. Nader did not make cars any safer at all. He did manage to sell a number of books though.
Hi SyncTronX,
Sorry, I have to completely disagree with you on the Nader thing. Man was not honest and most certainly in Ford's pocket.
Otherwise, I wish you success in obtaining test equipment. You might have it easier finding Tek 2213 or Tek 2235. The power supply is an easy fix in most cases. The conversation you are having with John should be in it's own thread or off-line (private).
Hi John,
You were not talking to me, but I am talking to you. You know better, but you are purposely going off topic about as much as you can. Talk about fixing test equipment. Looks like I'll have to request some assistance in cleaning this place since I am participating in this thread.
-Chris
I completely agree, although I'm a believer that at least one instrument should be known to be in tolerance and the readings should be transferable to other instruments. People who are skilled at calibration can do more with less. So it is the beginner that really needs to confirm the instruments are telling the truth.
Hi bsgd,
I do have your email (just checked). I am having a rough time at the moment, but I do intend to respond once I'm feeling up to it. The HP 339A is better than some people may give it credit for. You made an excellent purchase. Did you down load the service manual for it yet?
Hi Joshua,
In my opinion, Ralph Nader was a headline seeking so and so. There were plenty of FORD products busy taking lives due to cost accounting measures, and yet he never spoke up once about that. Methinks he was under an agreement with FORD (a four letter word). I owned a couple F products that were actually dangerous, the Pinto was the best known. But then, all those cars where the transmission slipped out of park to kill people. Then there were all the gas line clamps that sprayed fuel and started fires.
My parents owned a pair of Corvairs (one was a Monza) that were excellent cars. The problem was that the weight was in the rear and people spun out do to a lack of skill coupled with poor driving decisions. Bad you say? Talk to me about a Porsche or VW bug.
The cars Ralphy boy killed were safer than the ones he allowed to take lives. I don't know how well known it is, but GM did more work in crash studies than anyone in North America. They implemented safety measures consistently before anyone else did on this rock. My Pinto truly was a poorly built POS that liked to drop it's gas tank on the road when it got hit from the rear (mine was). Luckily, the car was both out of gas and stolen when it got nailed.
You should study these "heros" before accepting what they apparently stood for. Nader did not make cars any safer at all. He did manage to sell a number of books though.
Hi SyncTronX,
Sorry, I have to completely disagree with you on the Nader thing. Man was not honest and most certainly in Ford's pocket.
Otherwise, I wish you success in obtaining test equipment. You might have it easier finding Tek 2213 or Tek 2235. The power supply is an easy fix in most cases. The conversation you are having with John should be in it's own thread or off-line (private).
Hi John,
You were not talking to me, but I am talking to you. You know better, but you are purposely going off topic about as much as you can. Talk about fixing test equipment. Looks like I'll have to request some assistance in cleaning this place since I am participating in this thread.
-Chris
Hi Morten,
The two things that will limit the range of a THD meter (assuming a much better source) are the residual noise level. Measure that with shorted inputs. The other is the depth of the null you can achieve. The sharpness of that null will yield more accurate numbers as that more accurately measures the level minus the fundamental.
I seem to remember some conversations between us regards your instrumentation on your bench. It works, and it's the best you have. Besides, if your sine source is noisy, it's pointless to go for the best THD meter.
-Chris
The answers you are getting are accurate within the accuracy and noise floor of that THD meter. At the time, it was one of the best out there. THD meters are essentially voltmeters. If that function is accurate, then the readings for distortion should be. Just look at the residual to ensure that the fundamental is removed.
The two things that will limit the range of a THD meter (assuming a much better source) are the residual noise level. Measure that with shorted inputs. The other is the depth of the null you can achieve. The sharpness of that null will yield more accurate numbers as that more accurately measures the level minus the fundamental.
I seem to remember some conversations between us regards your instrumentation on your bench. It works, and it's the best you have. Besides, if your sine source is noisy, it's pointless to go for the best THD meter.
-Chris
anatech said:
Hi Chris!
Thank you for responding. Yes, I did download the manual, but Im still not sure if getting the 339A was a good purchase....it looks like mine is defective, and I have absolutely no experience in fixing these kind of instruments, although Im willing to try it (I already followed the calibration steps and calibrated my unit for gain). But, even after the gain calibration, I still have issues... I used DeOxit on the contacts but still have problems.
Im runnin out of ideas hehehe. My unit still cannot measure voltage right. As for the distortion part, I dont know how to test it for accuracy as I dont have any other way of measuring distortion to compare....
I have probably bought more used boat anchors off eBay and other sources than anyone here. I have also collected more distortion analyzers than anyone else here. Too almost completely pointless pursuits. I have learned a few things.
1) Don't buy stuff that is recent unless you are certain that it works. it seems anything post 7K or 5K from Tektronix and post about 1995 from HP is only repairable at a "module" level. That means parts are unavailable and modules if you can get them will be expensive. The Tekscopes group on Yahoo is a very good source for info. There is similar for HP but more scattered about the web.
2) Portable scopes are harder to fix and less flexible than the benchtop plug in scope. For audio a 5K Tek is a really good value. The 5440 is a 50 MHz scope and there are a plethora of plugins available for less than the shipping will cost. 7K scopes are also very good value with many plugins for specialized applications. I have a sampling plugin with a 30 pS risetime and a high gain one with a 50 uV/div sensitivity. And I have an absolute limit of $300 for any purchase. This illustrates that you can get close to the current state of the art in many ways for not much investment. And if something pops from old age another is not expensive.
3) The HP 330 and 330b are really too old to be too useful. The HP 331-334 series are really solid state versions of the 330, with auto nulling. They were near state of the art in the late 60's and early 70's but I would go for a Sound Technology analyzer at a minimum. Having both generator and analyzer together helps and they are really a value now (annoying since I have three and really don't need any). Other good candidates are Krohn Hite analyzers that have a good autonull and autoranging. The Ambers are also good but the small ones are hard to work on, they are very densly packed. The The Radiometer sweeping analyzer is really pretty neat but hard to find and really complex. Its best suited to tape and disk applications. The HP339 is also a good candidate. The knobs seem to be fragile judging from the number on eBay with broken knobs. The HP 8903 series are OK but are harder to work on and don't seem to respond to tweaks. The early ones (8903A) have a pseudo balanced input and won't work with a balanced output amp. The Tek AA501 or AA5000 are really first class and adequate to almost any current task. The Boonton 1120 (and 1121, there isn't a real difference between them) is what I use now. It has GPIB control so I can automate measurements, something I find very useful. If I need really low distortion I get out the ShibaSoku. Its large and a little clumsy to use. I have two of the matching oscillators. The combination has a distortion residual below .0001%, very nice but I don't think its that important anymore.
For new audio analyzers there are tree choices- all quite expensive. Rohde and Schwartz has a very nice system with low residuals and lots of features. AP has an extensive line with high performance and lots of options but the one I would get is the Stanford Research SR1
SR1 I think its the best value and its a fresh new design, as well as pretty cheap for this class of instrument.
Back on track improving any analyzer from the ST onward is worthwhile, the older ones really aren't. And a good sound card for $150 and good software + a signal conditioning box to protect it will get you more info than any of these boxes stand alone will. Think through why you are getting a distortion analyzer or a scope. They are not ends unto themselves, just tools.
1) Don't buy stuff that is recent unless you are certain that it works. it seems anything post 7K or 5K from Tektronix and post about 1995 from HP is only repairable at a "module" level. That means parts are unavailable and modules if you can get them will be expensive. The Tekscopes group on Yahoo is a very good source for info. There is similar for HP but more scattered about the web.
2) Portable scopes are harder to fix and less flexible than the benchtop plug in scope. For audio a 5K Tek is a really good value. The 5440 is a 50 MHz scope and there are a plethora of plugins available for less than the shipping will cost. 7K scopes are also very good value with many plugins for specialized applications. I have a sampling plugin with a 30 pS risetime and a high gain one with a 50 uV/div sensitivity. And I have an absolute limit of $300 for any purchase. This illustrates that you can get close to the current state of the art in many ways for not much investment. And if something pops from old age another is not expensive.
3) The HP 330 and 330b are really too old to be too useful. The HP 331-334 series are really solid state versions of the 330, with auto nulling. They were near state of the art in the late 60's and early 70's but I would go for a Sound Technology analyzer at a minimum. Having both generator and analyzer together helps and they are really a value now (annoying since I have three and really don't need any). Other good candidates are Krohn Hite analyzers that have a good autonull and autoranging. The Ambers are also good but the small ones are hard to work on, they are very densly packed. The The Radiometer sweeping analyzer is really pretty neat but hard to find and really complex. Its best suited to tape and disk applications. The HP339 is also a good candidate. The knobs seem to be fragile judging from the number on eBay with broken knobs. The HP 8903 series are OK but are harder to work on and don't seem to respond to tweaks. The early ones (8903A) have a pseudo balanced input and won't work with a balanced output amp. The Tek AA501 or AA5000 are really first class and adequate to almost any current task. The Boonton 1120 (and 1121, there isn't a real difference between them) is what I use now. It has GPIB control so I can automate measurements, something I find very useful. If I need really low distortion I get out the ShibaSoku. Its large and a little clumsy to use. I have two of the matching oscillators. The combination has a distortion residual below .0001%, very nice but I don't think its that important anymore.
For new audio analyzers there are tree choices- all quite expensive. Rohde and Schwartz has a very nice system with low residuals and lots of features. AP has an extensive line with high performance and lots of options but the one I would get is the Stanford Research SR1
SR1 I think its the best value and its a fresh new design, as well as pretty cheap for this class of instrument.
Back on track improving any analyzer from the ST onward is worthwhile, the older ones really aren't. And a good sound card for $150 and good software + a signal conditioning box to protect it will get you more info than any of these boxes stand alone will. Think through why you are getting a distortion analyzer or a scope. They are not ends unto themselves, just tools.
bsgd said:
We need some more detailed symptoms to figure out what the 339 needs.
First, what do you mean "My unit still cannot measure voltage right."? Is it high, low or something else? And what are you referencing to to determine that? Second, do you have a scope? Sometime a look at a waveform will tell a lot. As to distortion there are adjustments inside that you can optimize if you have a low distortion source. If the internal source is working well it would be OK for that task.
I don't remember the details, but you can calibrate a THD analyzer with just a resistor or two, and a signal generator. The procedure is in the construction article for the Cordell analyzer on his web site. Basically you knock the signal down to a low level and set the frequency to a harmonic (2nd?), and read it. As said above, a voltmeter is the heart of it.
Mr. Curl, I for one certainly "get it" about starting off with something cheap and upgrading to better as one can afford too, rather than spending weeks or months trying to convert tin stuff into golden stuff.... I've got plenty of used test equipment, can't recall anything I ever bought new. So making old gold shine, well that seems reasonable to me.
For instance, I like my HP-339A, have two of them, broken knobs and all. I got rid of the ST-1700A that was forever cranky, itchy and scratchy! I have like six backup scopes, not even counting the 7630a or the Fluke I recapped and left in the closet, but the scope I use just won't fail (B&K-1570A). I also have a B&K 2190, with Ch1 blown ($15 on ebay five years ago) due to a cheap NEC dual Fet that hasn't been made in fifteen years, thinking of trying an LSK389c in its place when I can find the time.
My ST-1500A is only partially working, been scratching my head on that on and off for over three years. I could send it out to CA. for a $995. flat fee fix plus $200. in both ways shipping costs, but I have more time than money, so that's the answer there.
Sure I'd like to be able to measure -125dB SN/R on my 339A if I could, now it gets to -112dB / -115dB one unit does a little bit better than the other. Maybe if I pulled the 1000 uF caps and put in 3,300 uf caps before those LM325H used as Vregs in the PS??? As for the 339A's generator output IC's from 30 years back, whatever they are... won't the LM4562NA be at least 10 dB better in terms of noise and distortion? Maybe put both 4562 chips in a dual IC in parallel to get even better performance?
Demian - I'd love to have a Rohde and Schwartz, played with it, took photo's of it at a Javits Convention Center show a few years back, but the 30,000usd that it costs is more than I made in the last 15 years total - gross huh?
-Steven
For instance, I like my HP-339A, have two of them, broken knobs and all. I got rid of the ST-1700A that was forever cranky, itchy and scratchy! I have like six backup scopes, not even counting the 7630a or the Fluke I recapped and left in the closet, but the scope I use just won't fail (B&K-1570A). I also have a B&K 2190, with Ch1 blown ($15 on ebay five years ago) due to a cheap NEC dual Fet that hasn't been made in fifteen years, thinking of trying an LSK389c in its place when I can find the time.
My ST-1500A is only partially working, been scratching my head on that on and off for over three years. I could send it out to CA. for a $995. flat fee fix plus $200. in both ways shipping costs, but I have more time than money, so that's the answer there.
Sure I'd like to be able to measure -125dB SN/R on my 339A if I could, now it gets to -112dB / -115dB one unit does a little bit better than the other. Maybe if I pulled the 1000 uF caps and put in 3,300 uf caps before those LM325H used as Vregs in the PS??? As for the 339A's generator output IC's from 30 years back, whatever they are... won't the LM4562NA be at least 10 dB better in terms of noise and distortion? Maybe put both 4562 chips in a dual IC in parallel to get even better performance?
Demian - I'd love to have a Rohde and Schwartz, played with it, took photo's of it at a Javits Convention Center show a few years back, but the 30,000usd that it costs is more than I made in the last 15 years total - gross huh?
-Steven
john curl said:
john curl said:
Conrad Hoffman said:
The international standard is ISO10012 (Management systems of measurement)
Conrad Hoffman said:
Yes can be done, but the values should be followed according to the manual or service manual.
The accuracy of the multimeter in general is 1%-1,5% (portable multimeters) this error rate is an error of scale (SPAM)if you calibrate for a different value of the manual, will have an error of scale.
Conrad Hoffman said:
Or ICs generator, build with resistors and capacitors of good precision, the accuracy of the ICs is reasonably for many oscilloscope, for oscilloscope expensive and sophisticated makes no sense to have good instrument without calibration appropriate, better to use a laboratory certified.
I have an ST1510 and the manual. Getting it working wasn't too difficult, only an output opamp was blown. They can be a real bear to fix if the problem is deeper. Figuring out how its intended to be used is another chore.
The ST1700 stuff is all cranky. Thats why I moved on. The R&S looks great but the SR1 seems to do all of the same things and more with its advanced digital analysis of a digital data stream. And its only $7K. Ouch.
The ST1700 stuff is all cranky. Thats why I moved on. The R&S looks great but the SR1 seems to do all of the same things and more with its advanced digital analysis of a digital data stream. And its only $7K. Ouch.
HP-339A problems
bsgd -
Inside your HP339A there are two pairs of +/-15 volt supplies in there, one pair for the OSC, one pair for the THD section, if either of these is seriously off, or has ripple or hum due to failed parts then nothing that follows will work right, check these lines with a scope/Dmm first.
-Steven
bsgd -
Inside your HP339A there are two pairs of +/-15 volt supplies in there, one pair for the OSC, one pair for the THD section, if either of these is seriously off, or has ripple or hum due to failed parts then nothing that follows will work right, check these lines with a scope/Dmm first.
-Steven
bsgd said:
Another possibility use a notebook or desktop PC with a Scope/FFT software; there is a free beta software that works really well that I have on my P4 Notebook, google Visual Analyzer 10.0.5 software - A powerful software to implement a software Spectrum Analyzer, Oscilloscope, frequency meter, distorsiometer, Volt meter and more; runs on Windows 9x,ME,2k,XP,NT,Server,Vista,7, or Linux using the wine software.
-Steven
-Steven
The new modular analyzer system from LinearX looks interesting:
http://www.linearx.com/products/analyzers/LX700MAIN/LX700_01.htm
Pricing seems reasonable.
Lukas
http://www.linearx.com/products/analyzers/LX700MAIN/LX700_01.htm
Pricing seems reasonable.
Lukas
LukasLouw said:
Yes looks quite interesting. Although, for stereo use you would need two source and two analysis modules.
I am also intrigued by this bit:
/quote
Discrete LA717 Opamp
The first input stage or front end of any spectrum analyzer is one of the most critical stages. While most other analyzers utilize monolithic commercial opamps, the LX721 employs a discrete opamp, the LA717. This discrete opamp contains over 80 components in less than 1 sq Inch. Many monolithic opamps have low voltage noise but have high current noise. The LA717 features both extremely low voltage, current, and flicker noise by using super high transconductance JFETs. The LA717 noise voltage is 0.95nV/RtHz and current noise is less than 0.08pA/RtHz. This results in up to 8dB noise improvement over monolithic opamp front ends through the front end 100K attenuator. The LA717 features computer matched JFETs, 30V/uS slew rate, 50MHz gain bandwidth product, 97dB loop gain, 75mA output current, and ±16mA supply current.
/end quote
Jan Didden
janneman said:
Yes, I calculate a mainframe, 2 generators & 2 analyzers comes to US$3,875.
Considering I paid near US$12,000 for my Audio Precision System One, analog only with a few options such as A weighting filter, IMD, burst & noise, about 13 years ago, this sounds like a deal.
mmm I wonder what I can get for the System One nowadays.....
Lukas
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.