and now you can wrestle with the unanswerable question "how do I quantify goodness-of-fit" ?
Numerical analysis textbooks typically say
Badness ( = 1/Goodness ) := Sum_over_all_datapoints_j { ( Unit1_measurement[j] - Unit2_measurement[j] ) ^2 }
I.e. the sum of the squared errors. However many people, including me, prefer the sum of the squared percentage errors :
Badness := Sum_over_all_datapoints_j { [ (meas1[j] - meas2[j]) / meas1[j] ] ^2 }
because, those people believe, a 10% error at low current is just as bad as a 10% error at high current.
Numerical analysis textbooks typically say
Badness ( = 1/Goodness ) := Sum_over_all_datapoints_j { ( Unit1_measurement[j] - Unit2_measurement[j] ) ^2 }
I.e. the sum of the squared errors. However many people, including me, prefer the sum of the squared percentage errors :
Badness := Sum_over_all_datapoints_j { [ (meas1[j] - meas2[j]) / meas1[j] ] ^2 }
because, those people believe, a 10% error at low current is just as bad as a 10% error at high current.
The horrible 2022 semiconductor shortage..... We are trying best to recovery production.
If there software guys, and any similar project, please PM, it think it would be best to have the energy one one single project, so a great achievement can be reached.
I have software issue here, the friend who are in charge of software have to leave for a while, bad news.
Although I can just release and sell out the stock I have, I don't think this is a good idea.
This tracer should be kept developed and keep improving its capability so every DIYers are happy.
Profits aren't important, I just want to have a tracer that at least meet my standards.
Not exactly sure what you are asking, but it sounds a bit like you should publish your code so people could pick this up and continue development.
Dear mbrennwa, at the moment your PyPSU CT is the best bench for a hw and sw developer. The best idea around.
It would help if you stated more formally what software skills you're looking for, development environments & what work remains to be completed before V1 can be considered ready.
For example I can do c/c++/c#/java but not Python.
For example I can do c/c++/c#/java but not Python.
Portability so it can work on both Windows and Linux would be preferable, and keep it open source too.
Yes, I noticed that you have suggested to publish/open the code more than once, thank you for suggesting. I have replied your suggestion after discussing with the author, who designed the whole tracer for both software and hardware, owning the full intellectual property of the product.
I do appreciate your kindness of open-sourcing your tracer project, that's a great contribution to the DIY community.
However, I think you can't ask people to do the same especially the guy who are not the owner of the design at ALL
In this case, I CANNOT open or publish things that doesn't belong to me; trust me, I also hope to open the code too!
I have formally asked the author and got the answer "Sorry, NO" firmly,
I respect his decision, and I have to , because it's not mine.
As probably explained before, I am not a coding/software guy, not even a bit. I don't make contribution to the coding work, and I don't own the property of the software program as well. I just made some suggesting to the UI and language translation of the program, and helping the author for communicating with the Western world.
I know it's good to open the code, making the software more powerful and versatile, but again, please respect the author's decision.
BUT the hardware is open, at least the schematic is open, the best results of my negotiation.
I hope I have explained the reason clearly so that you will not ask for opening again, though I wish to open too.
Yes, sure,
About what I am asking for or talking about, is that
Can the community together reconstruct and develop a new and open-sourced(as you and I want) tracer system on the basis of iTracer?
My idea is resulted from the fact that , as the author is busy and seems to be slowing down the software maintenance,
given the open hardware some other open-sourced designs, plus the manuals explaining the mechanism of working,
CAN the community work together to make an open-sourced version of the iTracer?
I am not software guy, but I can help to layout, and make a new iteration to fulfill the new system.
If this is possible, and everyone is contributing their specialties of this new open-source system,
including both software and hardware, then we can make a project that are free, open, user-friendly, and inexpensive.
DIYers can build up their own iTracer-like system, and with lowered costs, it could be the common language that DIYers uses to talk about their DUTs, sharing data, building Spice models, so many!
This is what I am talking about, but since I have no software background, I am sorry I am unable to state formally what software skills is needed, but I guess, if we are develop software based on hardware, maybe the software language is not limited specifically?
Now, we have open hardware that has been tested and proved to be working well, and I have got extra 5 programmed MCUs for testing and evaluation; I believe all these would be helpful for the open-iTracer, if possible.
An
Is there open available hardware? Are there open available schematics?
Last edited:
In principle, that's surely possible. In realitity, it would be most helpful if the protocol for communication with the iTracer hardware was fully documentented. Is it?
The hardware is closed too. Or do you plan to release at least the Gerbers?
Why don’t you repeat to the community what you replied to my inquire? You will not be able to order empty PCBs nor will you get Gerbers to produce your own PCBs even if you buy our MCU…
Last edited:
You totally don't get what I am saying.
Empty PCBs is from the author, and I replied you only assembled PCBs is available for sell, this is not decided by me, you don't fully understand the threads and the position I am at.
And now we are talking about is to make another iteration of iTracer due to the shortage and making it open.
By the way, I really don't like your manner in the PM and here.
Last edited:
I repeat what I sent you before. I wish you good luck with your project!
Very professional way to communicate with potential customers I must say.
Last edited:
You seriously asking people to contribute to a complete closed source project? The software is closed, the firmware is closed the schematics are released but the PCBs are closed source. If people contribute what is the return? A MCU with a closed firmware where you are not being able to even root out bugs? You mentioned the software developer is busy doing other stuff, so, I invest a lot of work and end up with a half ready product not being able to fix?
No thank you, there are other ways to waste time…
P.s. Unfortunately another Locky-project.
Sure, I hope so, but the protocol is also not available to me.
My plan now is to make a new layout of PCB first, at least solving the problem of PCB shortage.
But right now I just want to say forget my asking help for software, perhaps it is better to call for help after I've finished the PCB.
I didn't expect to see so many problems happened and got my idea distorted, but anyway, thank you for replying and caring about this, deeply appreciated. 🙂
Also, for people following this thread, I have found something interesting and helpful, another decent project, as I've introduced on another post, and I am putting it here again for reference. It's the tracer design by Dr. Borgmann, the following contents briefly introduce this outstanding design.
Check it out here:
http://diy.ucborgmann.de/index.php/en/audio-messtechnik-2/characteristic-curve-recorder
Specs in brief:
Vg: +-12V
Ig: +-120mA
Vd: +-20V
Id: +-2A
Simply speaking, the tracer has the ability up to 20V/2A, with up to 16Bit ADC used.
This spec is great enough for amateur DIYers, satisfying most measurements of audio transistors, although probably not fully enough for output-stage power transistors.
Mechanism of Working Diagram:
Amazingly, the DAC could be upgrade from 12bits to 16bits! (With LTC1655 module, and independent Vref=2.5V)
The available sources are below, including schematic, GERBER, BOM, MCU FW and Windows software.
Assembled tracer looks like this:
NOTICE:
I noticed there are fly-wires of the PCB, I am not sure if it is corrected in the GERBER file, so make decision and take the risk by self.
I am not responsible if resources available from the Website which may or may not be correct, accurate and complete.
Any replication and usage of the resources from the Website is at your own risk, as also state above.
All credits belong to the website's author.
Check it out here:
http://diy.ucborgmann.de/index.php/en/audio-messtechnik-2/characteristic-curve-recorder
Specs in brief:
Vg: +-12V
Ig: +-120mA
Vd: +-20V
Id: +-2A
Simply speaking, the tracer has the ability up to 20V/2A, with up to 16Bit ADC used.
This spec is great enough for amateur DIYers, satisfying most measurements of audio transistors, although probably not fully enough for output-stage power transistors.
Mechanism of Working Diagram:
Amazingly, the DAC could be upgrade from 12bits to 16bits! (With LTC1655 module, and independent Vref=2.5V)
The available sources are below, including schematic, GERBER, BOM, MCU FW and Windows software.
Assembled tracer looks like this:
NOTICE:
I noticed there are fly-wires of the PCB, I am not sure if it is corrected in the GERBER file, so make decision and take the risk by self.
I am not responsible if resources available from the Website which may or may not be correct, accurate and complete.
Any replication and usage of the resources from the Website is at your own risk, as also state above.
All credits belong to the website's author.
@jensenhoffman
If possible, Please avoid using pictures with transparent background such as the 2nd one in your post #119
It's very difficult to read for those of us who are trying to save our eyes by setting the forum theme color to Dark mode, please see following link below, thanks.
https://www.diyaudio.com/community/...rallel-instead-of-bridged.391167/post-7146386
ps. thanks for your effort in trying to realize the curve tracer.
If possible, Please avoid using pictures with transparent background such as the 2nd one in your post #119
It's very difficult to read for those of us who are trying to save our eyes by setting the forum theme color to Dark mode, please see following link below, thanks.
https://www.diyaudio.com/community/...rallel-instead-of-bridged.391167/post-7146386
ps. thanks for your effort in trying to realize the curve tracer.