Folks,
I am truly humbled by your interest in my THD analyzer, after about 30 years. I still use it myself partly because I'm just comfortable with it. I did it simply because I could not afford a good commercial one back at the time. Indeed, even now, I cannot bring myself to spend the big bucks on an AP - and I envy those who have one.
I own an HP339 analyzer (I got it for free) and actually prefer to use mine.
I did some minor upgrades about a year or two ago and I'll try to recollect those. I think one of them was to replace the output average detector with an RMS detector. Another was to put trim pots on all of the JFET control elements to optimize distortion reduction in them (I found that the usual 50% gate feedback is not always exactly optimal).
This biggest challenge with the original analyzer was the rats nest of multipole switches, which would be quite expensive these days. A multiplicity of relays and their associated passive tuning elements down on the PCB would make more sense and greatly reduce assembly labor.
I'd probably tweak the design of the SV oscillator a bit in accordance with some of the things people did on the wonderful audio oscillator thread, and some of those tweaks would likely be applicable to the SV notch in the analyzer section. I could be tempted to lower some of the impedances in the SV tuning circuits to get the noise down a little more.
Of course, adding digital meters and maybe a frequency counter and frequency trim and maybe a PIC processor would also be tempting if they could be done carefully without impacting the noise performance. An external power supply in a separate box would also not be out of the question.
Balanced inputs and outputs would make a lot of sense, too.
Many of these things would not add a lot of cost to the analyzer.
It would also make sense to put it all on a single board and use dual op amps to save some space. Ultimately, a surface-mount-implemented design would make a lot of sense, but would not be as DIY-friendly.
Not everyone needs the increased performance afforded by the tracking residual filters, so they could be optional and replaced by the usual fixed filtering choices on commercial analyzers. The 80kHz filters on commercial analyzers are a joke for THD-20 measurements.
The analyzer could also be equipped with an optionally-engaged class D measurement filter, like the AES-17 filter, but maybe not as complex. I'd probably go for a 7th order Bessel to keep a linear phase characteristic rather than the brick-wall design of the AES-17. All one needs to do is keep the spurious ultrasonic signals from a class D amplifier small enough so that they do not degrade the performance of the analyzer circuits.
Of course, all of this needs to be put in the context of the PC-based analyzer arrangements based on sound cards that are now available. What they can do is truly impressive, and building the necessary interface box like Pete Millet's is not difficult. If need be, I think there is room for improvement in that design, but for most it is great.
Cheers,
Bob
Hi Bob,
The PIC controller I'm using with my SVO seems to have the least impact on the oscillator.
They run a a such a high frequency any junk is way out of band. It's more challenging getting USB sound cards to work cleanly with them. That's where they become interfering. Ground isolation becomes necessary. Just for tuning and switching range they are not any more interfering than any other method.
The PIC controller I'm using with my SVO seems to have the least impact on the oscillator.
They run a a such a high frequency any junk is way out of band. It's more challenging getting USB sound cards to work cleanly with them. That's where they become interfering. Ground isolation becomes necessary. Just for tuning and switching range they are not any more interfering than any other method.
Last edited:
I still use mine and find it a superbly designed instrument, wanting for almost nothing. Though I use PC-based instruments, there's something comforting about doing a reality check with the Cordell. If the PC stuff isn't set up just right, the results can be off by a mile. With the Cordell I can look at the residual signal and know what's going on. And happy retirement Bob!
Thanks for your very kind words, Conrad.
Tomorrow is my last day. I really look forward to getting out of the rat race and spending more time with Angela (and audio), but I will also miss many great friends from work.
But I have many great friends here
.Cheers,
Bob
I have a mouthwatering desire to create a version of this with an LCD display and keypad for selecting the options all controlled by a microcontroller(interfaced with optofets for resistor selecting). It would be a surface mount version with larger components (resisters and capacitors would be no smaller than 0805 packages) so it would step up the difficulty but not be impossible. This would be a large and serious project for DIY'er but I believe I could handle it since I did graduate from an electronic engineering technology program and have job experience in surface mount soldering, PCB design with Altium (couldn't be able to afford it for personal use though), and a good interest in audio and filter fundementals. For me the largest difficulty would be in the programming required for the microcontroller, LCD, and keypad interfacing. As well as preventing high frequency clock noise from the digital portions of the project entering the analog sections.
How on earth would I be able to balance this with an 8-5 job, my other hobbies, and a social life though?
How on earth would I be able to balance this with an 8-5 job, my other hobbies, and a social life though?
I have a mouthwatering desire to create a version of this with an LCD display and keypad for selecting the options all controlled by a microcontroller(interfaced with optofets for resistor selecting). It would be a surface mount version with larger components (resisters and capacitors would be no smaller than 0805 packages) so it would step up the difficulty but not be impossible. This would be a large and serious project for DIY'er but I believe I could handle it since I did graduate from an electronic engineering technology program and have job experience in surface mount soldering, PCB design with Altium (couldn't be able to afford it for personal use though), and a good interest in audio and filter fundementals. For me the largest difficulty would be in the programming required for the microcontroller, LCD, and keypad interfacing. As well as preventing high frequency clock noise from the digital portions of the project entering the analog sections.
How on earth would I be able to balance this with an 8-5 job, my other hobbies, and a social life though?
If you want to use the Microchip stuff I can help you with that.
It's probably the most supported micro controller out there and tons of stuff on the web.
Free tools and 32 bit processing. Offers a small free DPS library as well.
A single-board version with frequency-switching relays on the board instead of rotary switches would be a good approach, even if a uc was not used. Also, use of dual SMT op amps and 0805 SMT resistors would help make it practical to get it all on one single PCB. The biggest PIA and cost factor in the original design was all of the rotary switches and related wiring.
Cheers,
Bob
Cheers,
Bob
Davada -- Your excellent generator design (which you so generously allowed me to see) is well under way also. It should see the light of day before too long. I hope we get to have a pcb available also.
So many good designs are starting to gather interest.... we need great low distortion sources to go along with the THD and FFT sound card etc and other gear to measure our audio projects and designs. SIM are important part of design but cant replace the test equipment for measuring the built design.
Thx-RNMarsh
Last edited:
If I ever get my PCB from China. It seems to take forever but the price is right. It still needs to be pruned before being ready for DIY and I need a post amp that can keep up. Then an attenuator. I,m also testing replacing the AGC with a micro controller.
After that I can contribute to Bob's design. I think to satisfy all a desing should be scalable.
Reducing the size of the power supply would also be great for using a smaller enclosure. There are some regulator circuit designs could be used to support DC or AC external supplies without line noise being an issue I believe. That's one area where I need to study more though haha
From a hobbyist and DIY perspective - and perhaps even gaining a few dB of performance - I'd like to keep the generator section on a separate PWB from the measuring section.
Dale
I do like the idea of an external power supply for both space and noise reasons, as long as it would not be too inconvenient for some to have essentially two boxes. In one of my preamps I use an external wall-transformer-powered +/- 20V regulated power supply which I then re-regulate to +/-15V in the preamp itself. That works quite well.
AC never sees the inside of the preamp. The re-regulation references the +/-15V internal supplies to the local preamp circuit ground.
Cheers,
Bob
If the design could fit into one of these enclosures: Box Enclosures - Product
..I honestly wouldn't mind the extra brick from the external supply (which probably wouldn't have to be that big I hope). I think 6 inches for an enclosure let alone the board might not be enough for the amount of components though, especially if I am to add the microcontroller interfacing. But hey might as well create the schematics and annotate to a pcb layout to see what really happens.
This is starting to coagulate in the back of my mind during work which means I should really consider putting some time aside haha
..I honestly wouldn't mind the extra brick from the external supply (which probably wouldn't have to be that big I hope). I think 6 inches for an enclosure let alone the board might not be enough for the amount of components though, especially if I am to add the microcontroller interfacing. But hey might as well create the schematics and annotate to a pcb layout to see what really happens.
This is starting to coagulate in the back of my mind during work which means I should really consider putting some time aside haha
Hi,
Is there anyone out there who made Giulianodes version of Bob's Distortion Analyzer?
I had a brief contact with Giuliano and unfortunately he did not yet finish his creation.
I must say he did a great job in re-designing the PCB layout.
I got the Gerber files of the layouts and now I am trying to figure out how to connect them.
Comparing the layouts with the photos posted in this threat I noticed some differences.
The AD536 is not present anymore on the meter PCB and the resistors like R(E) and R(F) plus relays PCB is not provided (but I guess it's easier to mount them on the rotary switch).
My idea is to follow Bob's design with the PCBs of Giuliano. I want to split the sine generator and analyzer meter into two cabinets. It does not fit in the housings I already have (first I wanted to build a Hameg HM8027, but after comparing specs with Bob's design I said goodbye to that idea).
Futhermore I want to use a digital meter i.s.o. the analog meter.
A PIC to drive the relays and a tekst display is nice, but too difficult for me (never learned myself to write programs).
At the moment I'm drawing new schematics with the relays in. Will take some time before I finish that.
BR, Rob.
Is there anyone out there who made Giulianodes version of Bob's Distortion Analyzer?
I had a brief contact with Giuliano and unfortunately he did not yet finish his creation.
I must say he did a great job in re-designing the PCB layout.
I got the Gerber files of the layouts and now I am trying to figure out how to connect them.
Comparing the layouts with the photos posted in this threat I noticed some differences.
The AD536 is not present anymore on the meter PCB and the resistors like R(E) and R(F) plus relays PCB is not provided (but I guess it's easier to mount them on the rotary switch).
My idea is to follow Bob's design with the PCBs of Giuliano. I want to split the sine generator and analyzer meter into two cabinets. It does not fit in the housings I already have (first I wanted to build a Hameg HM8027, but after comparing specs with Bob's design I said goodbye to that idea).
Futhermore I want to use a digital meter i.s.o. the analog meter.
A PIC to drive the relays and a tekst display is nice, but too difficult for me (never learned myself to write programs).
At the moment I'm drawing new schematics with the relays in. Will take some time before I finish that.
BR, Rob.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.