Tried using my function generator set to square wave and it probably isn't producing a perfect square wave and I get 45.4 on the keithley and 44 on the hp, so not too much difference but definitely some.
@@ptronix Oh, gosh. I'd expect both of them to read a bit low since it's hard to make a perfect square wave. 45.4 vs 44 doesn't really give much of a clue though.
Hi. Some suggestions. Use amplitude adjustments with an pot in the - op amp feedback path. Now its inverting output but that is no a big deal. Amplitude adjustments with an potentiometer as an attenuator is bad in such designs. Even if you have a load on it. Adjust r4 with an pot in parallel that it just starts to oscillate. Measure the resistor value and insert a new one. Use fkp2 wima 10nf and metallfilm resistors. Try an opa2156 or try other bicycle lamps. Output of osc should give about 2vrms. Keep up the good work.
Thanks for your interesting suggestions. I'm not going to do anything else with it until I get a new opamp. I've ordered opa4134 which has quoted figures of 0.00008 % compared with the tlo74 of 0.003% Also I tried various bulbs and the best was a very small "grain of wheat" type with very low current through it. I think the circuit is as good as I can get it, but will be interesting to see how the new opamp behaves
@@ptronix Sounds good. Have the fine tuning of R4 in mind. it's very important for the lowest possible distortion to have a gain of close to 1. That means the osc just begins to oscillate at 1khz. I should take some sec to start. Great op amp.
That's a bit difficult. Its really hard to pin down the specs for the 2015. What it does say is that connecting the internal source to the input should give around 0.2 % which seems a little high to me. Forgive me I can't find specs for the hp 8903 at the moment as the manual is several hundred pages!
Hi Nick, people already commented a lot even what I would have said. I watched with interest as I found myself stuck in the loop of different readings by different instruments. I do agree with you, the more instruments you have the more confusing it gets. As I’m sure you know it all goes back to calibration to a standard reference and then tolerances in the accuracy. I was buying second hand and old instruments so at some point I decided to put a point and bought a couple new with certificate of calibration. Based on those I aligned mines. Nice little project your pure 1KHz tone generator. I guess it would be handy if you can vary the output level. About my yaesu transceiver I didn’t find good power transistor yet, just crap from china. Unfortunately now I cracked a tunable transformer coil in the TX path and I’m stuck there as I couldn’t find the replacement yet. Without that the radio is useless. Hope all the rest is well. Here is hot like hell. Can’t afford anymore this heat. Best wishes Diego
Hi diego, nice to hear from you. I think when I'm trying to measure such low levels of distortion as I am you're doomed to failure with the old equipment I'm using. But I'm still quite impressed with what I've achieved with a relatively simple circuit. But to be fair to the keithley and the hp they do tend to agree when measuring levels in the 10's of %. Will be interesting to see if the exotic opamp I've ordered makes any improvement. But due to the UK being a third world country now I've been waiting over a week for it as the postal service gets worse and worse. Shame about your radio, all you can do is keep looking for parts I guess. I've cracked the cores of transformers before but managed to get the broken bits out and replace the core, but certainly not easy. Well autumn is not far off now so hopefully your high temperatures will decline. In fact we have winds from the north here so getting a lot cooler. But July and August have been fairly awful, a lot of rain and wind compared to last year. But I think I'd rather have that than the extreme heat southern Europe has had. Best wishes nick
Hey there... nice update video. Carl Jung must be right in his synchronicity theory, because it's not the first time I see someone working on a subject exactly on the same time period as me. It was exactly yesterday I was going through - again after some years - Rod Elliot's article (from Elliot Sound Products) on wien bridge and low distortion oscilators (did you read his article? It's good material for thoughts!). My little 2 cents on your dillema... 1. On Measuring instruments... It's really a rabbit hole when we start venturing into the realm of almost unmeasurable quantities. Measuring 0.004% or 0.0011% of any quantity requires top notch metrology grade equipment and pretty much anything can swamp your measurements. Even dissimilar metals in the connectors can start generating thermal noise that could ruin your day. But your situation is not that bad as, even though your instruments don't agree in numbers, they agree in showing which one is better, which one is worse. You'd never know what is the actual figure, but if you're chasing lowest possible 'anything' all you need is a compass. I would investigate though, on the keithley side, there must be something wrong there because keithley don't usually screw up so badly as an order of magnitude as you're seeing. 2. On the generator itself, the catch point is the operating point of the lamp... If it's running so low to be influenced by the ambient temperature, then it's not operating at its region where it actually self regulates the gain. Conversely, if you lower the positive feedback path resistance you increase the overall amplitude (which is good in many senses) and bring the lamp closer to an operating point where it's actually regulating the gain... but... Now you're reaching the limit of current one can expect to draw from an opamp without overstressing it (and overstressing it WILL increase its distortion. At this low levels of distortion you must really find an optimim balance. The best will be always with higher voltage and lower current lamps. That's where it enters Jim Williams choice for the #327, because it's a 28V lamp. By the way, I found a source of such lamps... They're still used widescread in aircraft button illumination, so, any aircraft mechanic will usually have tons of them. As a different path, I remember having seen a link to a design (if my memory doesn't fail, in Elliot Sound Product's page) of a guy who found a way to do the stabilisation "solid state", avoiding the lamp altogether, but it was a fairly complex idea... He used some sample and holding to sample the peak amplitude and provide a stable control voltage to the control element, which I dont remember right now if it was a FET or an LDR. Anyway. I leave you with those comments and look forward into seeing your progress.
Hello thank you for that long and interesting comment. I've not read the article by Rod Elliot but have now found his website and notice there is one on low distortion sine wave oscillators which I will read. Yes I agree I am on the road to nowhere trying to measure such low levels of distortion with the equipment I have available which is old and not been calibrated for many years. The Keithley looks incredibly difficult to calibrate compared to the HP so I'm not even going to go there. But on the other hand they do tend to agree if the distortion is 10's of %. In fact I was reading a data sheet for a low distortion opamp and it said it is virtually impossible to measure such low levels! Anyway its been an interesting project and good that I've managed to get such low levels that I've measured. I'm still waiting for a better opamp but as the UK is now a third world country the post takes ages! but I wouldn't be surprised if there is no improvement. The lamp I'm using is I think a 12v "grain of wheat type" I tried several and that was the best. It draws around 20ma at 6 v. I also tried changing the resistor that feeds it and 330 ohm seemed to be optimum. So look out for the next part when I get my exotic opamp! Best wishes Nick
Nick, good to see an update on your Quest for Ultra-:Low THD. Watching with a lot of interest. It makes me wonder - does the measuring device with the lowest reported distortion mean it's the most accurate?
Hi Jo, interesting question, which I don't know the answer to. I guess you would need reference standard of distortion if such a thing exists to check both the instruments. But they're both fairly old so there's bound to be errors I think. But when I've done measurements at higher levels of distortion I think the readings were fairly close. Anyway I will be able to notice any improvement in my circuit when I get my posh opamp. How's the psu?
@@ptronix PSU repair keeps getting interrupted by other tasks, however I did finally find some DIP clips (handiest tool for instrumenting ICs!) so I can now compare behaviors between the two channels right at the pins of the 723s. Fortunately both of those regulators are accessible while the unit is in-box and powered up. The op amps are another story - they are out of pin-probing reach due to the construction of the box. None of the parts in the critical path get hot. Surprised that I don't have part numbers tattooed on my fingertips given the number of times I've used the 'manual' temperature sensing approach. Can't instrument a thing once I pull out the board, of course, because the cabling just won't reach the board. At least it appears now that I don't need to drill out any rivets, as the rivets go to a strip of steel that can be unbolted from the box (according to what I can see). -Joe
As far as I'm aware both the keithley & the hp have high impedance inputs. So the only termination is the 5 k level control in my box. I will try something like 600 ohms and see if that affects the readings
Actually, I meant the other devices. I have the same HP as yours. it requires termination to have an accurate HD reading. I have that Keithley also. However, If I recall correctly, you can set the termination via the soft buttons.
Yes it's fairly easy to adjust frequency it's a matter of changing two resistor and two capacitors you could even use a potentiometer to make it variable. But I wanted a fixed frequency of 1 kHz
One simple thing you could try in order to do a sanity check on the meters is to feed them a pure square wave. A pure square wave has a THD of 48.3%.
Tried using my function generator set to square wave and it probably isn't producing a perfect square wave and I get 45.4 on the keithley and 44 on the hp, so not too much difference but definitely some.
@@ptronix Oh, gosh. I'd expect both of them to read a bit low since it's hard to make a perfect square wave. 45.4 vs 44 doesn't really give much of a clue though.
Hi. Some suggestions. Use amplitude adjustments with an pot in the - op amp feedback path. Now its inverting output but that is no a big deal. Amplitude adjustments with an potentiometer as an attenuator is bad in such designs. Even if you have a load on it.
Adjust r4 with an pot in parallel that it just starts to oscillate. Measure the resistor value and insert a new one. Use fkp2 wima 10nf and metallfilm resistors. Try an opa2156 or try other bicycle lamps. Output of osc should give about 2vrms.
Keep up the good work.
Thanks for your interesting suggestions. I'm not going to do anything else with it until I get a new opamp. I've ordered opa4134 which has quoted figures of 0.00008 % compared with the tlo74 of 0.003%
Also I tried various bulbs and the best was a very small "grain of wheat" type with very low current through it.
I think the circuit is as good as I can get it, but will be interesting to see how the new opamp behaves
@@ptronix Sounds good. Have the fine tuning of R4 in mind. it's very important for the lowest possible distortion to have a gain of close to 1. That means the osc just begins to oscillate at 1khz. I should take some sec to start. Great op amp.
Have you read the spec sheets for both instruments to find what their level of uncertainty is?
That's a bit difficult. Its really hard to pin down the specs for the 2015. What it does say is that connecting the internal source to the input should give around 0.2 % which seems a little high to me. Forgive me I can't find specs for the hp 8903 at the moment as the manual is several hundred pages!
Hi Nick, people already commented a lot even what I would have said.
I watched with interest as I found myself stuck in the loop of different readings by different instruments. I do agree with you, the more instruments you have the more confusing it gets. As I’m sure you know it all goes back to calibration to a standard reference and then tolerances in the accuracy. I was buying second hand and old instruments so at some point I decided to put a point and bought a couple new with certificate of calibration. Based on those I aligned mines.
Nice little project your pure 1KHz tone generator. I guess it would be handy if you can vary the output level.
About my yaesu transceiver I didn’t find good power transistor yet, just crap from china.
Unfortunately now I cracked a tunable transformer coil in the TX path and I’m stuck there as I couldn’t find the replacement yet. Without that the radio is useless.
Hope all the rest is well. Here is hot like hell. Can’t afford anymore this heat.
Best wishes
Diego
Hi diego, nice to hear from you. I think when I'm trying to measure such low levels of distortion as I am you're doomed to failure with the old equipment I'm using. But I'm still quite impressed with what I've achieved with a relatively simple circuit. But to be fair to the keithley and the hp they do tend to agree when measuring levels in the 10's of %.
Will be interesting to see if the exotic opamp I've ordered makes any improvement. But due to the UK being a third world country now I've been waiting over a week for it as the postal service gets worse and worse.
Shame about your radio, all you can do is keep looking for parts I guess. I've cracked the cores of transformers before but managed to get the broken bits out and replace the core, but certainly not easy.
Well autumn is not far off now so hopefully your high temperatures will decline. In fact we have winds from the north here so getting a lot cooler.
But July and August have been fairly awful, a lot of rain and wind compared to last year. But I think I'd rather have that than the extreme heat southern Europe has had.
Best wishes nick
Hey there... nice update video.
Carl Jung must be right in his synchronicity theory, because it's not the first time I see someone working on a subject exactly on the same time period as me.
It was exactly yesterday I was going through - again after some years - Rod Elliot's article (from Elliot Sound Products) on wien bridge and low distortion oscilators (did you read his article? It's good material for thoughts!).
My little 2 cents on your dillema...
1. On Measuring instruments... It's really a rabbit hole when we start venturing into the realm of almost unmeasurable quantities. Measuring 0.004% or 0.0011% of any quantity requires top notch metrology grade equipment and pretty much anything can swamp your measurements.
Even dissimilar metals in the connectors can start generating thermal noise that could ruin your day.
But your situation is not that bad as, even though your instruments don't agree in numbers, they agree in showing which one is better, which one is worse.
You'd never know what is the actual figure, but if you're chasing lowest possible 'anything' all you need is a compass.
I would investigate though, on the keithley side, there must be something wrong there because keithley don't usually screw up so badly as an order of magnitude as you're seeing.
2. On the generator itself, the catch point is the operating point of the lamp...
If it's running so low to be influenced by the ambient temperature, then it's not operating at its region where it actually self regulates the gain.
Conversely, if you lower the positive feedback path resistance you increase the overall amplitude (which is good in many senses) and bring the lamp closer to an operating point where it's actually regulating the gain... but... Now you're reaching the limit of current one can expect to draw from an opamp without overstressing it (and overstressing it WILL increase its distortion.
At this low levels of distortion you must really find an optimim balance.
The best will be always with higher voltage and lower current lamps.
That's where it enters Jim Williams choice for the #327, because it's a 28V lamp.
By the way, I found a source of such lamps...
They're still used widescread in aircraft button illumination, so, any aircraft mechanic will usually have tons of them.
As a different path, I remember having seen a link to a design (if my memory doesn't fail, in Elliot Sound Product's page) of a guy who found a way to do the stabilisation "solid state", avoiding the lamp altogether, but it was a fairly complex idea... He used some sample and holding to sample the peak amplitude and provide a stable control voltage to the control element, which I dont remember right now if it was a FET or an LDR.
Anyway. I leave you with those comments and look forward into seeing your progress.
Hello thank you for that long and interesting comment.
I've not read the article by Rod Elliot but have now found his website and notice there is one on low distortion sine wave oscillators which I will read.
Yes I agree I am on the road to nowhere trying to measure such low levels of distortion with the equipment I have available which is old and not been calibrated for many years. The Keithley looks incredibly difficult to calibrate compared to the HP so I'm not even going to go there.
But on the other hand they do tend to agree if the distortion is 10's of %.
In fact I was reading a data sheet for a low distortion opamp and it said it is virtually impossible to measure such low levels!
Anyway its been an interesting project and good that I've managed to get such low levels that I've measured.
I'm still waiting for a better opamp but as the UK is now a third world country the post takes ages! but I wouldn't be surprised if there is no improvement.
The lamp I'm using is I think a 12v "grain of wheat type" I tried several and that was the best. It draws around 20ma at 6 v. I also tried changing the resistor that feeds it and 330 ohm seemed to be optimum.
So look out for the next part when I get my exotic opamp!
Best wishes Nick
hi, can you please indicate where to find the article describing the oscillator? Thank you.
try this
www.elektormagazine.com/articles/reference-sine-wave-generator
@@ptronix is the correct link. Thank you
@@WalterBrollo just checked it works for me, just click on where it says download
@@ptronix yes I do 🙂
Nick, good to see an update on your Quest for Ultra-:Low THD. Watching with a lot of interest.
It makes me wonder - does the measuring device with the lowest reported distortion mean it's the most accurate?
Hi Jo, interesting question, which I don't know the answer to. I guess you would need reference standard of distortion if such a thing exists to check both the instruments. But they're both fairly old so there's bound to be errors I think. But when I've done measurements at higher levels of distortion I think the readings were fairly close. Anyway I will be able to notice any improvement in my circuit when I get my posh opamp. How's the psu?
@@ptronix PSU repair keeps getting interrupted by other tasks, however I did finally find some DIP clips (handiest tool for instrumenting ICs!) so I can now compare behaviors between the two channels right at the pins of the 723s. Fortunately both of those regulators are accessible while the unit is in-box and powered up. The op amps are another story - they are out of pin-probing reach due to the construction of the box.
None of the parts in the critical path get hot. Surprised that I don't have part numbers tattooed on my fingertips given the number of times I've used the 'manual' temperature sensing approach.
Can't instrument a thing once I pull out the board, of course, because the cabling just won't reach the board. At least it appears now that I don't need to drill out any rivets, as the rivets go to a strip of steel that can be unbolted from the box (according to what I can see).
-Joe
Are you terminating the signals correctly?
As far as I'm aware both the keithley & the hp have high impedance inputs. So the only termination is the 5 k level control in my box. I will try something like 600 ohms and see if that affects the readings
Hi, tried terminating with a 560 ohm resistor. Made no difference, so I'll wait to see what the new opamp achieves
Actually, I meant the other devices. I have the same HP as yours. it requires termination to have an accurate HD reading. I have that Keithley also. However, If I recall correctly, you can set the termination via the soft buttons.
Can u adjust the frequency? Or is it fixed?
Yes it's fairly easy to adjust frequency it's a matter of changing two resistor and two capacitors you could even use a potentiometer to make it variable. But I wanted a fixed frequency of 1 kHz