In order to set the 1khz to 0db I must lower the output watts to 178mW. I don’t think that is creating a proper test parameter:
One small step at a time:
I am still concerned with why the 1khz amplitude is above 9
I keep reading that this should not exceed 0
These performance test to verify the unit meets or exceeds original spec
The ideal scenario would be to have screen shots as I have posted above for these test
What “Simulated Test” conducts the individual test 1 - 13
You have said previously your amp should have a gain of 30. But it’s showing around 18 dB. Why is that?
I am still concerned with why the 1khz amplitude is above 9
I keep reading that this should not exceed 0
Your generator is at 6 dBV. Your gain is shown as 18.22 dB. Given those two items, wouldn’t you expect the output of the amp to be 24.22 dBV = 16.26Vrms = 66.1W. Everything that is being measured makes perfect sense, doesn’t it?
note if you set your Stop Level to be 6 dBV, that is the QA403 generator level. Your amp will add its own gain beyond that.
The ideal scenario would be to have screen shots as I have posted above for these test
What “Simulated Test” conducts the individual test 1 - 13
Yes, these are all doable. But not “push a button and measure it all automatically” as there is different operator setup needed between several of the test. A poster on the forum named @Var has made some great videos showing how he iterates through various tests and confirms vintage amp performance against published specs.
Gain 18.27 + 7.90 = 26.17
Is the amount of gain to drive to specified output
103 watts into 4 ohms
17.90 + 10 = 27.90 is the point of clipping
154 watts into 4 ohms
That 7.90 or 10 is passing through the Volume Control which is attenuating that value. Volume is set around level 7. Thus the amount of drive to the amp section is obviously less.
Hi @YamahaCA, ok thanks. Your earlier statement “This is a 70wpc integrated amp, 30db gain” is with volume at max? And the other plots you are showing is with volume at at some level other than max?
I think that generally when testing an amp you want to run it wide open at max volume and plot that.
The QA401 seems to have been developed primarily for design engineers, who regularly use such equipment. But because of its affordability, the unit is popular with vintage audio hobbyists, who might not be so experienced. In an effort to have a simplified step-by-step method of setting up the equipment and software, I have used the article posted under the QA401 Product Information area of Quantasylum.com called Basics: Power Amps, which discusses the setup of a 1000W power amp.
I find this article to be most helpful but the wording and screenshots reference the QA401 making it difficult to follow, if this was updated to reflect the UI of the QA403 it would be of a major benefit to all QA403 owners.
Should QA hire a Technical Writer it would be a major benefit to all users:
Although, I guess we should all feel privileged to have that 1950’s automotive users manual.
I find these feelings of the Users Manual in every direction that I seek assistance:
As I mentioned some time ago the manual for this thing is inscrutable. Don’t have one, don’t know how to run it… hopefully someone who has one and is able to decipher the manual will chime in.
The QA401 seems to have been developed primarily for design engineers, who regularly use such equipment. But because of its affordability, the unit is popular with vintage audio hobbyists, who might not be so experienced .
Hi @YamahaCA, unfortunately, this is kind of the nature of the beast with test equipment. You can purchase a $50,000 RF spectrum analyzer, and the manual won’t tell you a thing about how to measure the 1 dB compression point of an amp. It just assumes you know.
As another example, I have a $3000 oscilloscope. The manual doesn’t tell me anything about how to actually measure. It tells me how to not kill myself, it tells me what the various knobs do. But if I want to measure a DDR read/write operation, it’s not much help.
Technical writers can’t really help here, because they don’t know how to measure the 1 dB compression point of an amp or how to measure DDR memory either. At best, they can take a procedure from an engineer and convert it into a readable and pretty doc. But the engineer needs to first document the procedure. And that is very, very time consuming and there are a long list of these tests, and each test can point back to many standards. There are US standards, EU standards, Asia standards, and even among EU standards each country will often have its own standards. And then you have several audience types to consider: The first are design engineers that have done this before, they just need a little help moving to a new platform. The second are service techs that might be used to older equipment (such as HP8903 audio analyzer) and they need to be brought up to speed on the new equipment from an application perspective. And then you have very sophisticated hobbyists that have the time to invest and forge their own path. And then you have complete beginners that aren’t sure what gain and noise really mean.
To address the above, you really are looking at book-length levels of documentation.
The Big Change that is Coming
AI will play a transformative role here. Now, this doesn’t help you if you need to measure something TODAY. But take a look at this exchange with ChatGPT on how to measure amplifier distortion:
Note that the AI answer is pretty good. Vague, yes. But pretty good. Additionally, note that it happily codes up some custom code using PyQa40x The code isn’t quite correct. But I put PyQa40x on Github to see how the various AIs would learn from that. And they are learning…
And so, I think interactive places where information is exchanged between users (such as this forum) play a very important role informing AIs that can then (at some point) use that knowledge to inform users. Every code sample on the forum, every place one user explains to another user how something is done get inhaled by an AI
Evolving Products
When products didn’t evolve, the car analogy you bring up makes a lot of sense. And old test equipment used to have thousands of pages of documentation, including schematics, calibration and sophisticated troubleshooting flow charts. But my car software gets updated every month, and there’s no real manual. My phone is updated quarterly, and there’s no manual. I go to some links at Google if I have a question and then try to decipher why the android pages are all telling me something different.
For example, let’s say I want to turn off randomized mac addresses on my phone. Searching on google doesn’t even point me to a Alphabet/Android authority. They instead reference a web page called armstrongonewire as the top link on how do it. But that link is old and no longer applicable and the instructions are just flat out wrong. One of the world’s richest companies, with no authoritative source on how to use their product!
The point here is that products that are updated and evolve will always be behind on documentation. Things that are easy to understand might not need any documentation. Things that are not might get a wiki page. The manual is functional (“this button turns on A-weighting” without explaining what A-weighting is). The wiki pages are more like app notes (“Here’s why you might want A-weighting…”).
In any case, I hope this helps. I don’t think there will be a book explaining everything you might want to know about measuring audio with the QA40x anytime soon…come to the forum and ask if its not clear from the docs and github pages. That makes us all smarter (me because I didn’t know people were having trouble with that, and you because the docs weren’t clear). The interaction will inform other and AI, and that will be huge in the end.
And we’re not too far off from asking GPT/Claude/whatever “I have a McIntosh MC2300 and I’d like some code to sweep 1 kHz from 1W to 100W and plot that in svg in the top half of a page. and then some more code to sweep frequency response via chirp at 1W output and plot that in svg at the bottom half of the page. Title that as " Bill Johnson’s MC2300 After Repair”, and save as PDF. Also, if you could indicate the MC2300 specs per the manufacturer on the plot, that’d be great. I’m using a 4 ohm load"
The AIs are remarkably close to being able to do this…We just need to collectively keep shoveling them info…
@Matt- the point with the $50k spectrum analyzer and 1dB compression point hits the nail on the head. However, HP/Keysight/Agilent (or instance) may have staff develop AP Notes (and now videos) to do certain measurements using their expensive piece of gear. The “reference library” for the QA40x is certainly adequate to get started and if you run into something weird or need a little help, the forum is great as are you with supporting it. Just my 2cents
1 Like
Is there a way to attach a file here?
Let’s try this
Have you seen this before?
I don’t recall seeing that tutorial, but it looks pretty well done. When I get I chance I will look at it more closely, but he did a really nice job from what it looks like.
1 Like
Hopefully this will insert the missing link that I have been asking about.
Again, my request on how to present the test procedure in a user friendly format.
I spoke of the need to change the QA401 to the QA403 to make it easier to follow.
Yes, the doc is very nicely done.
Is this something you could provide without having to land on Mars first?
Then if the same format could be created for these tests:
Without having to send a maned mission to Uranus first?
@YamahaCA I believe I covered how I set up to do most of the measurements mentioned in the spreadsheet I gave you, which you have posted here, in the MAC2300 video. I would probably change a few things were I to re-do it, which I don’t plan to do. At some point I want to go through testing a preamp with the QA40x, but am not sure when that will happen…
1 Like