One og my 401k funds may carry a few of their shares, so I'm not so worried that they asre justvtrying to sell more opamps. As the wise man said, " They ain't in it for their health, kid". More annoying is no comment about setting the 500-ohm trimpot. Whgile it obviously (I think) optimises the feedback, I wouldn't know hoew to set it without a distortion analyser that has a THD+N floor lower than the circuit. This something I don't have. It would be nice to know if there is a voltage reading somewhere that would accomplish this.
At least I'm not a complete ignoramus: I've figured out two mods to make this more practical. One is trivial; just the observation that C1&C2 need to be switchable to get a useful range. Better still a 4-ganged rotary switch to provide a set of fixed frequencies. The second one is that the VTL5C10 is nearly unabtainable unless you want to buy in volume from P-E, but you can use two VLT5C4/2'a from Allied in parallel to get very neary the same voltage/current to resistance curve.
At least I'm not a complete ignoramus: I've figured out two mods to make this more practical. One is trivial; just the observation that C1&C2 need to be switchable to get a useful range. Better still a 4-ganged rotary switch to provide a set of fixed frequencies. The second one is that the VTL5C10 is nearly unabtainable unless you want to buy in volume from P-E, but you can use two VLT5C4/2'a from Allied in parallel to get very neary the same voltage/current to resistance curve.
Homebrew/Secondhand
This is an interesting discussion.
I've seen a lot of designs for instruments like frequency meters over the years, and I'd always thought it might be a useful thing to have. Eventually I bought a secondhand Racal for £50. Overnised xtal, 200 MHz and more features than you can shake a stick at. The magazine designs would have cost that much for parts and the end result would never have been as good. They don't use overnised xtals for a start.
I've just built a low distortion double integrator loop audio oscillator from a design in a defunct British magazine called ETI.
It's amplitude control is a precision full wave rectifier and LED/LDR. The article claims 0.001% THD at 1KHz and 0.008% at 100Hz.
After much fiddling with shielding and earthing, an HP8903 shows 0.009% and 0.015%. A Marconi TF2104, which costs about £100, does <0.003% in both cases. With just a bit more fiddling/trying different op amps (presently NE5534s, TL052s in the article) I'm convinced I'll get there. Of course, I'm using a crummy dual pot, not a nice one like you find in commercial instruments.
Was building the oscillator entertaining?
Definitely, and much better value than the equivalent expense in say cinema tickets or restaurants.
Was it a cost effective way of obtaining a low THD oscillator?
No.
Of course, you can regard building test instruments as a worthwhile activity in its own right.
There's also the possibility of tweaking or regutting commercial instruments.
Conclusions:
The design and connecting the parts is just a part of the story - what you're paying for in a commercial unit is the debugging, the robustness and the very nice presentation.
I'm not tempted to try the Cordell THD analyser, but the article is a good read. Might just try the oscillator section one day.
The Linear Technology AN 43 final design is worth a go if the parts aren't too dear, and the design in AN 67 is interesting.
I'll pounce on a Marconi TF2104, or better, the HP equivalent, when I see one for a price I like.
Secondhand instruments can be a very good buy. Broken ones can be dirt cheap and easy to fix.
The age of kit companies is not over for nothing.
It would be interesting to see the circuit of the oscillator in the 8903B.
Pete.
This is an interesting discussion.
I've seen a lot of designs for instruments like frequency meters over the years, and I'd always thought it might be a useful thing to have. Eventually I bought a secondhand Racal for £50. Overnised xtal, 200 MHz and more features than you can shake a stick at. The magazine designs would have cost that much for parts and the end result would never have been as good. They don't use overnised xtals for a start.
I've just built a low distortion double integrator loop audio oscillator from a design in a defunct British magazine called ETI.
It's amplitude control is a precision full wave rectifier and LED/LDR. The article claims 0.001% THD at 1KHz and 0.008% at 100Hz.
After much fiddling with shielding and earthing, an HP8903 shows 0.009% and 0.015%. A Marconi TF2104, which costs about £100, does <0.003% in both cases. With just a bit more fiddling/trying different op amps (presently NE5534s, TL052s in the article) I'm convinced I'll get there. Of course, I'm using a crummy dual pot, not a nice one like you find in commercial instruments.
Was building the oscillator entertaining?
Definitely, and much better value than the equivalent expense in say cinema tickets or restaurants.
Was it a cost effective way of obtaining a low THD oscillator?
No.
Of course, you can regard building test instruments as a worthwhile activity in its own right.
There's also the possibility of tweaking or regutting commercial instruments.
Conclusions:
The design and connecting the parts is just a part of the story - what you're paying for in a commercial unit is the debugging, the robustness and the very nice presentation.
I'm not tempted to try the Cordell THD analyser, but the article is a good read. Might just try the oscillator section one day.
The Linear Technology AN 43 final design is worth a go if the parts aren't too dear, and the design in AN 67 is interesting.
I'll pounce on a Marconi TF2104, or better, the HP equivalent, when I see one for a price I like.
Secondhand instruments can be a very good buy. Broken ones can be dirt cheap and easy to fix.
The age of kit companies is not over for nothing.
It would be interesting to see the circuit of the oscillator in the 8903B.
Pete.
Hello Pete and all,
I totally agree with your comments.
Refurbished instruments' prices can be very attractive, and you may find a good 8903 ( -A, -B) in a fraction of the price when new.
With my 8903B I can obtain about -94dB (0.002%) T.H.D. (+N) above almost the whole audio band, and this instrument is more complete than a simple distorsion factor meter with its oscillator.
The service manual with the electrical schematics can be downloaded from the agilent site, the file is "08903-90062V2.pdf".
Marcello
I totally agree with your comments.
Refurbished instruments' prices can be very attractive, and you may find a good 8903 ( -A, -B) in a fraction of the price when new.
With my 8903B I can obtain about -94dB (0.002%) T.H.D. (+N) above almost the whole audio band, and this instrument is more complete than a simple distorsion factor meter with its oscillator.
The service manual with the electrical schematics can be downloaded from the agilent site, the file is "08903-90062V2.pdf".
Marcello
Since I obviously was the one who kicked this thread alive again,-- maybe I should join in...
I do have an old HP339, which I haven't used since my tape recorder days,- way back when...
I do have the service manual for it, tucked away somewhere ,- too well - can't find it yet.. but I think it will be a lot harder to modify this old unit to compete with even the newer DIY ones.... am also aware of Ian Hickmans article in E&WW...
I personally think that it easier to work with a THD meter in some applications, rather than spectrum analysers, even if I do agree that in the finals, the spectrum is probably of more interest than the actual THD number.
For the cost of e.g. the Cordell analyser, - or the Hickman version, I think a lot of us has most of the components in stock anyway, after a great number of years in DIY and professional electronics.
Even if Ebay is accessible for me to, the freight cost of sending something close up to the north Pole is somewhat preventive in many cases... ( I live 69.4 deg N-16.1 E !!
)
For the software analysers, - I think they fulfill a slightly different niche, and also has other uses, but what I asked for was peoples impressions in using the programs,. I do have some smaller freeware versions, but these are not up to measure, for very low THD work....
I do have an old HP339, which I haven't used since my tape recorder days,- way back when...
I do have the service manual for it, tucked away somewhere ,- too well - can't find it yet.. but I think it will be a lot harder to modify this old unit to compete with even the newer DIY ones.... am also aware of Ian Hickmans article in E&WW...
I personally think that it easier to work with a THD meter in some applications, rather than spectrum analysers, even if I do agree that in the finals, the spectrum is probably of more interest than the actual THD number.
For the cost of e.g. the Cordell analyser, - or the Hickman version, I think a lot of us has most of the components in stock anyway, after a great number of years in DIY and professional electronics.
Even if Ebay is accessible for me to, the freight cost of sending something close up to the north Pole is somewhat preventive in many cases... ( I live 69.4 deg N-16.1 E !!
)For the software analysers, - I think they fulfill a slightly different niche, and also has other uses, but what I asked for was peoples impressions in using the programs,. I do have some smaller freeware versions, but these are not up to measure, for very low THD work....
AuroraB said:
Where the spectrum analyzer is really useful is in looking at power supplies -- and line harmonics
I am still scratching my head trying to figure out how to use the FFT to measure phase margin -- the folks at Tektronix said I will have to become more proficient at MathCAD ! (you can do it automatically with their TDS7XXX series and a pulse generator.)
if you are looking for a disto analyzer
here's a Krohn-Hite 6900A on the BAY -- these are truly "hands-free" devices with auto null for both frequency selection and input -- I have a 6900B and it is much easier to work with than the HP339 I own -- of course you only have a 1kHz oscillator. The measurements of the KH and HP units match exactly FWIW.
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=2562641478&category=25423&rd=1
here's a Krohn-Hite 6900A on the BAY -- these are truly "hands-free" devices with auto null for both frequency selection and input -- I have a 6900B and it is much easier to work with than the HP339 I own -- of course you only have a 1kHz oscillator. The measurements of the KH and HP units match exactly FWIW.
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=2562641478&category=25423&rd=1
"here's a Krohn-Hite 6900A on the BAY"
Yup. Got one a while back. Just the thing for the less astute(like me) - real plug-n-play. It seems to be a bit unstable at points in the spectrum where you have to switch from one range to the other. But otherwise no problems.
You can download manuals including service at the K-H website
Yup. Got one a while back. Just the thing for the less astute(like me) - real plug-n-play. It seems to be a bit unstable at points in the spectrum where you have to switch from one range to the other. But otherwise no problems.
You can download manuals including service at the K-H website
I figured I would open the service manual and work though the adustment proceedures. However, they specify having a couple of ultra low (<.001%) signal generators available. These are not easy to come by cheaply so I'm in the process of trying to build the one in the LT1115 datasheet. I should hve the parts by the end of the month. If it works I'll see what I can do with the K-H.
Meanwhile I'll look at the manual again -- maybe I can improve the adjustment in the interim.
Meanwhile I'll look at the manual again -- maybe I can improve the adjustment in the interim.
I adjusted it using the Heathkit IG18 and an HP3336B Synthesizer and HP465A amplifier -- you don't need that high precision to adjust the A.C. voltmeter, adjusting the discriminators is a matter of ramping the input at different frequencies and making sure that the comparators latch. You do need a scope for looking at the lissajous figure to adjust the balance on the oscillator.sam9 said:
For the precision resistors -- just make these up from standard values.
Many years ago I build Bob Cordell's THD analyzer. Sold it when I got something else, but still miss it.
Worked beautifully, with very low residual, auto tuning, auto nulling, BW up to 100kHz IIRC. The design should be on www.cordellaudio.com. At the time, he had pcb's available, which I got. It is a fairly involved design mechanically as there are several stepped switched with multiple wafers, but as a DIY analyzer it can't be beat.
Jan Didden
Worked beautifully, with very low residual, auto tuning, auto nulling, BW up to 100kHz IIRC. The design should be on www.cordellaudio.com. At the time, he had pcb's available, which I got. It is a fairly involved design mechanically as there are several stepped switched with multiple wafers, but as a DIY analyzer it can't be beat.
Jan Didden
A 1000 Hz sine generator and an accompanying distorsion meter with sub ppm performances are described in these downloadable files :
Cyril Bateman s'site :
http://uk.geocities.com/cyrilb2@btinternet.com/downloads_4.html
Download Capacitor Sounds series1 papers:
"CapSound1.pdf"capsound1pdf 1.1MB.
"CapSound2.pdf"capsound2.pdf 1.5MB.
"CapSound3.pdf"capsound3.pdf 0.7MB.
"CapSound4.pdf"capsound4.pdf 0.7MB.
"CapSound5.pdf"capsound5.pdf 0.7MB.
"CapSound6.pdf"capsound6.pdf 1.2MB.
Cyril Bateman s'site :
http://uk.geocities.com/cyrilb2@btinternet.com/downloads_4.html
Download Capacitor Sounds series1 papers:
"CapSound1.pdf"capsound1pdf 1.1MB.
"CapSound2.pdf"capsound2.pdf 1.5MB.
"CapSound3.pdf"capsound3.pdf 0.7MB.
"CapSound4.pdf"capsound4.pdf 0.7MB.
"CapSound5.pdf"capsound5.pdf 0.7MB.
"CapSound6.pdf"capsound6.pdf 1.2MB.
ionomolo said:
I did. It's an extremely sensitive method (you can measure amplitudes of under 100nV buried in noise), but requires a dual phase model, with variable (frequency an amplitude) internal reference signal. The method is also very time consuming (at least for the old-ish Princeton Applied Research I got from EBay). Newer models may have some degree of automation, but are certainly very expensive...
Unless you have deep pockets and/or plan to measure distortions well under 1ppm, it doesn't make much sense to invest time and money in an (e.g.) Signal Recovery device that will conveniently do the job. For that kind of money you can buy a very good (and relatively new) spectrum analyzer that will serve you well for 99.99% of the audio applications.