Test Equipment Anonymous

I guess the Hameg frame with the function generator and distortion analyzer can be considered an audio analyzer, albeit not a particularly good one.

Right away on power up, the distortion analyzer's display didn't come up but the function generator's did. I remembered from looking over the manuals online that each of the bays in the Hameg frame have their own power supplies because the plugins can program the voltages on some of the rails and that could differ from bay to bay depending on what the plug-ins in each specify. I decided to swap the distortion analyzer and function generator to see if the problem followed the distortion analyzer or the rightmost bay and its associated power supplies were at fault. When I removed the distortion analyzer plug-in, I saw this great BCS Electronics sticker on the top. It looks like BCS Electronics is long gone from what I can tell by searching the internet. I didn't try sending a Telex to that number at the bottom but one magazine ad from 1985 does show them carrying Hameg equipment.







The distortion analyzer came up in the left bay so I started testing it out. I was able to tune in the frequency adjustment and set the level, then go for a THD+N measurement and it's overstating the distortion from my 8903B's signal generator.



On the other hand, it can't detect any distortion in the output of the QuantAsylum QA403 whose output is very clean.

I shut down the frame and swapped the plug-ins back in order to try out the function generator. Unfortunately, with the dead bay, going from the function generator to distortion analyzer isn't going to be possible right now. Instead, I hooked the function generator's output up to the 8903B and QA403.





Adjusted to 1 kHz according to the Hameg, about 1 V RMS according to the 8903B.



Change the measurement on the 8903B to distortion and wow. 0.818%



This is the Hameg function generator's sine output seen on the QA403 and it's awful. I actually started laughing at how bad it was. I realized later when I was showing the pictures to a friend that the pictures actually show a good example of how a sine can look perfectly good on a scope despite being loaded up with all kinds of distortion that's only visible on frequency domain equipment. I also realized I didn't test any of the other waveforms the function generator can produce either. I shut everything down and met the friend for dinner after seeing how bad the sine output of the function generator is without testing it further.

Right now, I'm on the fence between taking the cover off the frame and seeing if I can find what's wrong with the right bay or clearing the bench off and putting the Amber on it tomorrow.

Analogue function generators using a shaping circuit, never produce a particularly low distortion sine wave, which is needed for distortion measurement.

It's strange that the Hameg distortion meter correctly detected no distortion from the QA403 but significant distortion from the 8903B. I would have thought the 8903B would produce unmeasurably low THD as far as the Hameg was concerned.

I suspect a bad cable or bad contact somewhere.

Is the Hameg distortion meter automatic, or do you have to find and maintain the null manually?

Zenith wrote: ↑Wed Sep 25, 2024 9:56 am Analogue function generators using a shaping circuit, never produce a particularly low distortion sine wave, which is needed for distortion measurement.

I wondered about that, but such generators seem to have a "nipple" on the waveform, and that Hameg doesn't.

Whether or not -50dBc is "low distortion" depends on application. Either way it is a good illustration of the limitations of low-end scopes for frequency domain purposes.

tggzzz wrote: ↑Wed Sep 25, 2024 12:01 pm

Zenith wrote: ↑Wed Sep 25, 2024 9:56 am Analogue function generators using a shaping circuit, never produce a particularly low distortion sine wave, which is needed for distortion measurement.

I wondered about that, but such generators seem to have a "nipple" on the waveform, and that Hameg doesn't.

My Thandar TG501 (probably from the 70s) doesn't have the pip on the peaks of a sine wave, because I've just checked with a DSO. If anything it looks slightly flattened and a bit noisy.

The specification for the sine wave is "Less than 0.5% to 50kHz, less than 1% to 500kHz; all harmonics >30dB below fundamental on 1M range." On a scope it doesn't look like a great sine wave. I think the pip was particularly noticeable on some IC function generator chips, the 8038 and the XR2206.

tggzzz wrote: ↑Wed Sep 25, 2024 12:01 pm Whether or not -50dBc is "low distortion" depends on application. Either way it is a good illustration of the limitations of low-end scopes for frequency domain purposes.

-50dB is about 0.3% THD. For most distortion testing you'd want a source that went below the bottom level of the distortion meter, which I recall is 0.1% on my rubbishy old things, better ones go far lower. When I borrowed an HP8903E, a Philips PM5110 (Wien Bridge, thermistor stabilised) was less than 0.01% THD over most of the audio range, (it wasn't sold as an ultra low distortion sig gen) and an AWA G232, about half that at most. Those analog FGs were quite versatile and they would sweep, but they had limitations.

Quite believable.

My only thing like that is a Keithley 2015THD. It is a standard multimeter which contains an audio DDS, and can measure the THD. The DDS has a 0% THD, a power amp I used to have had 0.011%.

Zenith wrote: ↑Wed Sep 25, 2024 9:12 pm

The specification for the sine wave is "Less than 0.5% to 50kHz, less than 1% to 500kHz; all harmonics >30dB below fundamental on 1M range." On a scope it doesn't look like a great sine wave. I think the pip was particularly noticeable on some IC function generator chips, the 8038 and the XR2206.

Correct. The 8038 chip does have a small pip on the sine wave. I built a basic tester with that chip and any decent scope clearly shows it.

8038 has a number of distortion issues. Some you can trim out of the sine shaper but the key problem is there is a comparator inside it with some feedback and hysteresis and when it switches it goes with a thump. The tail current source is just a resistor so you get garbage on the output every time it flips one way (can't remember which way). Terrible little IC but competent if you don't give a shit

If you want a sine wave, a function generator is not the place to get it, although you can make them fairly good with a decent low pass filter on the output. Those are of course difficult to make tuneable which is why you get an "meh it'll do" job.

If I wanted a low distortion signal source it'd be a wien bridge oscillator with some non-linear feedback (light bulb or FET).

bd139 wrote: ↑Fri Sep 27, 2024 10:47 pm 8038 has a number of distortion issues. Some you can trim out of the sine shaper but the key problem is there is a comparator inside it with some feedback and hysteresis and when it switches it goes with a thump. The tail current source is just a resistor so you get garbage on the output every time it flips one way (can't remember which way). Terrible little IC but competent if you don't give a shit

If you want a sine wave, a function generator is not the place to get it, although you can make them fairly good with a decent low pass filter on the output. Those are of course difficult to make tuneable which is why you get an "meh it'll do" job.

If I wanted a low distortion signal source it'd be a wien bridge oscillator with some non-linear feedback (light bulb or FET).

I'd use a DDS.

I have a 8951-based 50MHz source, but I spotted a glitch on its output. Turned out that it has an internal comparator used to generate a square wave with 50% duty cycle from the filtered sine output. The glitch occurred when the comparator switched.I

The glitch was difficult to spot and much smaller than the 8038 nipple.

From the "just because I can" and "gotta use it sometime" departments...

DDS THS measured with my Keithley 2015THD, 1kHz up to 9th harmonic

2015THD source: -0.029% -71dBc

8951: 0.041% -67.5dBc

IsoTech 4MHz sine/square/triangle: 0.067% -63.5dBc

Interesting. Not terrible!

I will say I used whatever digital arb I had as an HF RF signal generator fine. You couldn't tune any harmonics from a receiver outside it which was pretty good with a 2uV MDS.

bd139 wrote: ↑Fri Sep 27, 2024 10:47 pm If I wanted a low distortion signal source it'd be a wien bridge oscillator with some non-linear feedback (light bulb or FET).

Yes. A special thermistor produced pretty low THD easily but they are no longer made. The better sig gens used a CdS cell and LED. There are some very elaborate feedback stabilisation schemes to drive the THD really low.

My distortion test set is a 3-bay TM500 frame with a SG505 oscillator and the AA501 analyser. The SG 505 is impressive -- in the audio pass band it's specified to 0,0008% THD. I've got the version with balanced output which is a really nice thing to have.

According to the manual, it's a "phase shift oscillator" built from three NE5534 opamps.

tggzzz wrote: ↑Wed Sep 25, 2024 9:41 pm Quite believable.

My only thing like that is a Keithley 2015THD. It is a standard multimeter which contains an audio DDS, and can measure the THD. The DDS has a 0% THD, a power amp I used to have had 0.011%.

The DDS has a 0% THD

A likely story.

Zenith wrote: ↑Sun Sep 29, 2024 10:10 pm

tggzzz wrote: ↑Wed Sep 25, 2024 9:41 pm Quite believable.

My only thing like that is a Keithley 2015THD. It is a standard multimeter which contains an audio DDS, and can measure the THD. The DDS has a 0% THD, a power amp I used to have had 0.011%.

The DDS has a 0% THD

A likely story.

Irritatingly, you're right.

However, my Digilent Analog Discovery is <0.01%, -90dBc. That's an advantage of a 14 bit DAC