1
Project logs / Re: Really universal soldering controller
Also MAX5035's frequency of operation is 2 times lower, which means considerably larger inductance and output capacitance, so I am not sure if the substitution is at all possible.
Messages
This section allows you to view all Messages made by this member. Note that you can only see Messages made in areas you currently have access to.
Messages - sparkybg
Also MAX5035's frequency of operation is 2 times lower, which means considerably larger inductance and output capacitance, so I am not sure if the substitution is at all possible.
2
Project logs / Re: Really universal soldering controller
3
Project logs / Re: Really universal soldering controller
sparkybg, you will be able to make a new design of the pcb? what kind of help will be needed?A new design on this PCB is only possible if full project was available, with proper links between PCB and schematics.
A new design of my PCB is of course possible, but it is at least a week working only on it. I don't have that much spare time.
...But I am happy to see that it works even that and works well on both channels as I can see!.
Umm, there is offset between channels. So it does not work well on both channels. Moreover, because of the offset, you don't know if thge calibration seems ok because of the offset, or because of the gain.
4
Project logs / Re: Really universal soldering controller
if properly separate the analog, digital line into layers. there is a chance that pcb will work correctly.
Yes, if the PCBs are rerooted properly, they will work. Smoe parts also needs relocating. For example, the calinbation trimmer.
You cannot place the parts on the PCB without thinking what will happen. For example, All the pards of the imnput amplifier must be as close to the amplifier as possible. Then, all the parts of the secondary amplifier, the gain and the offset chips, must again be close to each other. The analog power and analog ground of the MCU also must be routed properly, the reference voltage should be referenced to the analog ground as close to the final amplifier as possible.
...an so on.
The power lines themselves are too thin. Tha ground planes have several bottlenecks. All this creates voltage drops and offsets. Even 10mA current on wrong place can give 20+ degrees offset. Yes, it is nasty, but that's the way it is.
@minkok The bad design is not because the enclosure limitation. It can be made better with the same limitation. The bad design is due to lacking knowledge of what happens on the PCB and how the device works.
5
Project logs / Re: Really universal soldering controller
I do not agree with you. " EXTREMELY BAD PCB design" definitely not! As we can see the board works stable, does not have visible noise, and I cannot differ from your board except for the difference b/n the channels ( maybe bad PCB manufactured or a bad component) and definitely could be fixed, but it needs the proper tools and of course a time.
Yes it FEFINITELY IS! MIxing high-gain analog and digital on a whole PCB, not thinking how to root the power and the GND planes always gives results like this. Using all 4 planes on 4 layer board for GND is nonsense.
The creator of this board has done a nice game of sudoku, but definitely not a nice PCB design. The offset speaks for himself. And it will stay this way no matter who will make your PCB.
And I speak from experience - as I said, this costed me two iterations of my PCBs in order to eliminate such things.
It is your decision eigher to lose or not lose your time wioth this nonsense, but it will NEVER work good.
6
Project logs / Re: Really universal soldering controller
where can see?
Here:
https://github.com/bakanyako/UN5.2C-TINY
But it is made by the same person that made your boards, so I am sceptical about it's performance.
7
Project logs / Re: Really universal soldering controller
Also, there is 2 channel version. Again made by chinese guy.
8
Project logs / Re: Really universal soldering controller
This is the power plane. Same goes for GND plane, both ot 3-rd layer and outer layers - it is again sepearate for analog, digital, and current sources. This way no voltage drops dur to currents can happen, nor noise can be injected digital part to the analogue one, where the big gain amplifiers are.
Notice thet it is divided by 3 - one branch goes to analaog circuitry, one branch to digital circuitry, and one branch to current sources. Also, there is a separate poer lines for OLED in order noise froim it not to penetrate into analog circuitry. All branches meet at the connector's 3.3V/GND.
If this is not done, many thins like this on your boards may happen, even if the compomnents themselves are absolutely OK.
So, I's 99.99% sure it is PCB design flaw. Sorry.
P.S.: Also there is separate SGND line, coming directly from the shunt on the back board. I cannot see if it is present on ypur boards.
So please, when we discuss ceofficients/offsets, don't rely on this PCB, but on real 5.2C.
P.S. 2: It's even worse! Instead of using one of the middle layers for 3.3V, both middle layers are used for GND, and there is only tracers for power, and the power for the MCU is taken at it's center.
Sorry again, but this is EXTREMELY BAD PCB design. I don't thing you will be able to make it work, no matter what you do.
9
Project logs / Re: Really universal soldering controller
There are no clear net names, so I am continuing to be unable to comment on routing. Also, some of the electronics is changed - I suspect SENSEA is connected directly to VoutA and SENSEB directly to VOutB, there's no Ground, the ID is unclear, an so on, sn so on....
The only thing that I can say is that it is very chatoic, and there is no star separation of digital and analog parts. So yes, the one who made it obviously never had a job with hi-gain circuits. In such circuits you must thing wehere the electrons are going.
10
Project logs / Re: Really universal soldering controller
When you have a differentioal setup with gain of 750, and this precise current sources, the rooting of the PCB is critical. This costed me 2 versions of the PCB until I got 5.2,
11
Project logs / Re: Really universal soldering controller
result 0, nothing has changed Channel B: SenseB and Vout1- R: 978 - 980, after warming up for 20 minutes R: 975 - 976.
Channel A: SenseA and Vout1- R: 1000 - 1002 not change after 20 minutes of warming up.
p.s I begin to doubt that the problem is in the components, 8 sets of 5.2c pcb with such a mistake.
Strange. No sush thing on my Tindie board. Maybe you are making the same mistake on every board.
You may make a hi res photo of your front board to compare to mine.
@Doomedahab: Difference of upto 3-5 is normal. Usually it is around 2 or less. On my board it is less than 1.
12
Project logs / Re: Really universal soldering controller
1. checked all resistors 0.1%And what is the result?
2. replaced the transistor front Q18, Q19, Q21, Q22, back Q2, Q8, Q10, Q11.
3. replaced the diodes front D18, D19, D20, D21.
4. replaced the front U12, U13, U17, U18, U16.
I took all the details from the 5.2 pcb (SenseA and Vout1- = R: 1000 - 1002 / SenseB and Vout1- = R: 999 - 1001).
for tweezers, changing cartridges will not help, there will be an offset of 20 degrees between the channels.Make R70 0 ohm (short it with solder or wire) and the offset will be gone, if everything else is OK.
13
Project logs / Re: Really universal soldering controller
If 1% resistors (instead of 0.1%) are used, than I cannot comment anything. It is of VITAL importance for resistors on both current sources AND amplifiers (especially the input differentioal maplifier - U13, U18) to be 0.1%.
I am using a Tindie board as a reference. There is NO visible difference between channels, and both current sources are within 0.1% from the reference 382.81uA. And this is even with chinese opamps here and there, which, as we know, are not the best ones.
So, there is definitely a problem on your board. The curerent sources are only part of it (although your meter can also may be a bit wrong).
With 379.7uA on channel one, you should calibrate to 992 instead of 1000. BUt again, if your differential amplifier is not made of 0.1% resistors (R52, R53, R54, R62, R63, R65), there will be errors and offsets present, especially when differential input is used, and this is the case with C245. If the last amplifier (U17) is not made with 0.1% resistors (R49, R50, R60), then there may be an offset error.
14
Project logs / Re: Really universal soldering controller
The resistance of the contacts is 5-10 times higher, but this does not matter in our case, because the input resistors are 15k already.
There is still some available from Farnell, more available from Mouser and Digikey.
15
Project logs / Re: Really universal soldering controller
I am struggling to find ADG734BRUZ. Is there an alternative quad-channel analog switch I can use?Unfortunately, not.
Although ebay/china is not the best place to buy chips, since it is not available nowhere else, it may be worth trying:
https://www.ebay.com/itm/373153877450
measuring without resistor?Yes.
Channel A: SenseA and Vout1-
Channel B: SenseB and Vout1-
And, until you don't have entirely correct setup, talking about coefficients is irrelevant.