Probably didn’t give it a serious listen, but don’t remember it being anything special sound-wise. Was mainly just a restoration project to keep me and it is a nice looking integrated.
-
Feeling lucky? Check out the AK 24th Anniversary Giveaway thread and join in the fun!
Analog Discovery 2 test results for various tube amps
- Thread starter BmWr75
- Start date
JonL
Lunatic Member
Not surprised. The OPTs are tiny compared to those in a Dynaco SCA-35 (or ST-35), also an EL84 PP amp.
Allen Organ Model 20a monoblocks - restored for use in HiFi applications, gain mod by Kegger
Push/Pull 6L6GC topology
Push/Pull 6L6GC topology
Attachments
If there was an easy way to do this for you I would, but there isn't. To make a table I would have to manually read the values off the graph and write them down.
There are many videos on YouTube that will help you understand how to read line graphs. Here is one example:
When ringing is present on a square wave, it is indicating a peak in the frequency response. Depending on where the peak is, it may be fine and not need to be addressed, or it could be an indication of instability (in a feedback amp). Either way, it should be assessed and decided if further action is needed by the designer.
This is not correct. The lowest frequency present in a 100Hz square wave is 100Hz. It is therefore not possible to see attenuation using this waveform to excite an amplifier that exhibits flat response to lower frequencies. The appropriate tools to analyze response other than ringing (which requires a 10kHz square wave or higher) are swept sines or pink/white noise.
Jack
restorer-john
Addicted Member
Jack, the slope/tilt of the square wave (top and bottom) is an indicator of phase shift, which shows frequency response variations above AND below the fundamental you are injecting. That's why you use 100Hz for LF extension investigation. It will show you almost down to DC. 10kHz is for HF and stability investigation.
Rod Elliot explains:
quote: "This is one of the great benefits of squarewave tests - you can see variations both above and below the fundamental frequency, something that isn't possible with sinewave testing. If you test with a sinewave and want to see the effect at (say) 20Hz, you have to use a 20Hz sinewave. With a squarewave, you can see the effect of boost or cut at frequencies well below the test frequency. A squarewave is unique in that regard, and that's why they are such a useful test tool."
You only have to inject a 100Hz squarewave into an amplifier and play with the tone controls to see exactly what I am talking about. In fact you can set an amplifier with tone controls to its most flat response in a few seconds with two square waves. As with all the various tube amps in this excellent thread, they are transformer coupled, so LF extension is limited and a LF squarewave will tilt down left to right all things being equal.
Last edited:
The tilt is indeed the result of phase shift, something I have mentioned here numerous times. But as is true concerning most aspects of square wave use, determining frequency response from this secondary characteristic creates inaccuracies because the frequencies necessary to a direct measurement aren't present. Measuring with sines produces the simplest and most precise result.
Jack
Your square wave shots consistently show a pronounced tilt at 1kHz. I've built and restored a lot of tube amps and I've rarely seen this kind of behavior, except maybe on some stubbornly out-of-spec units. Ringing aside, I find it hard to believe that a restored ST-70, or even Nick's 300B amps couldn't produce a flat square wave at 1kHz. Are you sure your equipment is properly calibrated?
No. I’ve investigated calibration of the AD2 and not found a way to do so using the BNC breakout board. Would you interpret this square wave as tilted? Just curious.
Attachments
No, that looks flat, though the ringing is something fierce. You can't connect the output of the signal generator to the input of the board? I suspect it's not the AD2 but the way it's connected to your computer, or maybe the probes you're using. I'm not trying to be snarky, but every single DUT here has a pronounced tilt at 1kHz, and that's doesn't seem right to me. It's as if you had a treble control set off-center. It calls into question the frequency response and distortion measurements.
I've ordered a BNC splitter cable to loop the output back to the input of the BNC board. Thought I'd already ordered the correct cable, but NO. Also found in the user manual how to calibrate frequency response using this loop back cable.
Okay, good! I don't mean to rain on the parade, it's a great idea for a thread, but it just seemed like something was consistently off.
Nothing wrong with pointing out potential issues when you see them. Accuracy is important, and it might save someone else a lot of grief.
Jack
Did the Frequency Response calibration today by looping back the AD2 BNC output to the BNC inputs by using a BNC splitter cable. Shown below are two frequency response charts for an Analog Ethos Legendarium 2A3 SET amp. I can't see the calibration made much difference in the FR curve shape.
