Page 3 of 13

Re: CJ125 board

Posted: Sat Feb 04, 2017 3:17 am
by kb1gtt
All PWM's should be fine as far as I can tell. You would probably get better heating with the faster PWM. The slower PWM would likely have more fatigue as the heater gets to heat and cool more. I would think this would create more long term fatigue. I would expect less hot / cooling with the faster PWM. However that's just a guess and for testing purposes all PWM's should be fine as far as I know. Also more switching means more heat dissipated in the ECU. However this MOSFET has a fast switching time, so the transition time is very small, so probably not much heating from this transitions state.

Re: CJ125 board

Posted: Sat Feb 04, 2017 12:45 pm
by AndreyB
And voltage divider was the problem! Without it the firmware was reading twice the voltage it was expecting, with the divider we have cold mosfet and warm o2 sensor :)
5) RESOLVED IN R0.3 Correct C9 should match C2 for 1kHz low pass filter.
6) RESOLVED IN R0.3 R12 is not routed right
7) voltage divider on UR and UA channels
Success!
[video][/video]

Re: CJ125 board

Posted: Sat Feb 04, 2017 4:13 pm
by puff
congrats! where do we find the source code?
just wondering, how do you set up PWM if you don't read datasheets? :D
Снимок экрана 2017-02-04 в 19.11.12.png
Снимок экрана 2017-02-04 в 19.11.12.png (80.8 KiB) Viewed 24608 times

Re: CJ125 board

Posted: Sat Feb 04, 2017 4:28 pm
by kb1gtt
Is this intended for Frankenso add-on board? Do we need a divider, or could we route that to an AN input on the frankenso board?

Re: CJ125 board

Posted: Sat Feb 04, 2017 4:54 pm
by puff
From my point of view there will be much more sense if it remains a standalone device: you'd need a relay to power it up and one analog input to read the data. If it were an add-on board, it would require five more wires for interconnect and actually no benefits.

Re: CJ125 board

Posted: Sat Feb 04, 2017 10:21 pm
by AndreyB
Is there value in producing a 3v signal right on the board? This case board more universal and we can route right to adc pin?

Re: CJ125 board

Posted: Sun Feb 05, 2017 12:20 am
by kb1gtt
To me it appears that UA should be fine, with no divider. If your lambda amplification is 17, a voltage of 3 is over 2.25. If we are above say 2.8V, I would suggest changing the "Initialisation register1" register "VL" bit to an amplification of 8 instead of 17. In general, we are probably fine with leaving the amplification at 8, as we are not emissions. We don't really need to be that tightly controlled. If we leave the VL bit at 0, for an amplification of 8, we would loose some accuracy in the readings, but would have the full range of lambda readings up to 6.0. However a lambda of say 1, might really be something like 1.1 or 0.9. So any how, I believe we would be fine with this as is with no divider.

About UR. This op-amps has an internal 15k ohm resistor. So if we put an external 21.6k for now soldered on top of C2, it would drop a 5V signal to 2.951 V. But of course who knows what the internal resistance tolerance is. Also it appears that this is expected to be calibrated. So it might need a software adjustment.

As far as I can tell, R8 and R15 should be changed to 0R.

Re: CJ125 board

Posted: Fri Mar 10, 2017 3:28 pm
by Rhinoman
I have a PCB, the CJ125 and the micro board, can I get the firmware please?

Re: CJ125 board

Posted: Sat Mar 11, 2017 1:40 pm
by AndreyB
Zero progress was made since on rusEfi cj125 code - need more code contributors.

You would need to poke @ for his firmware.

Re: CJ125 board

Posted: Sat Mar 25, 2017 11:47 am
by nighthunter
Hi, a working CJ125 driver would be great. Now i am using the hardware/firmware from elektor 1/2012 but it seems to have slow response, so tuning is a little bit tricky.

Re: CJ125 board

Posted: Sat Mar 25, 2017 12:49 pm
by AndreyB
nighthunter wrote:Hi, a working CJ125 driver would be great.
Accepting patches and monetary donations :)

Re: CJ125 board

Posted: Sat Mar 25, 2017 10:11 pm
by nighthunter
Accepting patches and monetary donations
Are you using paypal? Ill make you some donation next week. This is a good project.

Re: CJ125 board

Posted: Sat Mar 25, 2017 11:24 pm
by AndreyB
Top right corner of this page has paypal donate button :)

Re: CJ125 board

Posted: Thu Apr 06, 2017 6:03 pm
by nighthunter
Hi, are there any news about integrating wideband?

Re: CJ125 board

Posted: Tue Apr 18, 2017 5:19 pm
by AndreyB
nighthunter wrote:Hi, are there any news about integrating wideband?
I doubt I will have much time to work on it within the next few months. In light of http://rusefi.com/forum/viewtopic.php?f=4&t=1215 @ is our best hope at the moment :)

Re: CJ125 board

Posted: Mon Jan 22, 2018 5:26 pm
by AndreyB
@ believes that the following fixes are needed on the board:

Re: CJ125 board

Posted: Tue Jan 23, 2018 1:30 am
by kb1gtt
@ is correct as far as I know. It seems I had made an R0.3, but no checked it in. I'm not sure why. Perhaps we were still working on it and I got side tracked.

Re: CJ125 board

Posted: Tue Jan 23, 2018 3:28 am
by AndreyB
Sweet, just merged schematics changes and I see no recent PCB changes and the board already says 0.3

Do you think you've commited PCB changes a while back already and PCB and schematics are now in sync?

I will probably wait for @ to make more progress before ordering a fresh board.

Re: CJ125 board

Posted: Mon Jan 29, 2018 7:40 pm
by AndreyB
@ has implemented CJ125 control and we now have a build-in WBO controller! That's actually a pretty big deal :)



There is an opinion that

Code: Select all

 7)                  voltage divider on UR and UA channels, 5V --> 3V (not clear if we need this on the little board - we can always use external divider)
 8)                  Remove R3 (connect ends) and remove C3
since CJ125 is measuring nernst cell resistance between UN and VM

Re: CJ125 board

Posted: Mon Jan 29, 2018 8:06 pm
by Rhinoman
I don't think that you need to delete those components - its a fairly high frequency filter and the constant current source will maintain the correct current, maybe leave the positions and fit a 0R resistor if not required. From what I recall they are shown in another document.

Re: CJ125 board

Posted: Tue Jan 30, 2018 1:49 am
by kb1gtt
This is good news, and good job @

Agreed, R3 is now 0R, and C3 is not DNP. I left them for an optional filter. I see some of the OEM documentation shows this not detailed filter.

I do not understand UR and UA comment. Is that trying to say we should add a resistor divider such that the UR and UA signals are always below 3V?

Re: CJ125 board

Posted: Tue Jan 30, 2018 1:52 am
by kb1gtt
Also was that LSU4.2, or LSU4.9? Did Q1 start getting hot, or did that stay nice and cold?

Re: CJ125 board

Posted: Tue Jan 30, 2018 2:42 am
by AndreyB
kb1gtt wrote:
Tue Jan 30, 2018 1:49 am
I do not understand UR and UA comment. Is that trying to say we should add a resistor divider such that the UR and UA signals are always below 3V?
UA requires divider but looks like we are fine without a divider on UR. Next question is where should UA divider be located - if the little CJ125 board is designed for Frankenso or Frankenstein we have dividers there. If we want the little board to be self-contained for example to be used with cheap stm32f103c8 eBay board it would be a different story. I suggest we do not think about self-contained yet.

Right now I would plan our CJ125 board to be designed for Frankenso. I think we need to plan a specific location on Frankenso. I would like to suggest mounting flat below the LCD screen. In order to make things easier, we need to match GND/3v/5v pinout with the board, and we also need to add a few holes with 0.1" spacing just for mounting. I hope these pictures help to relay my proposal.

So

Code: Select all

 9)                  enlarge silkscreen font - current font is too small
10)                  match Frankenso GND/3v/5v pinout and provide 3 extra holes in corners just for flat mounting

Re: CJ125 board

Posted: Tue Jan 30, 2018 3:42 am
by kb1gtt
I'm thinking up by U6, as we have 12V there. I'm not sure how we would get the SPI lines connected. Perhaps we could add some vias on R0.6, and for now solder in some short jumpers over to a the TCPL pads. We have some analog lines which are near by as well. We could solder in jumpers for now, and again install via's for R0.6 later on. Hmmm, while I guess R0.6 could include this board, and it would not need to be integrated. I guess I'm also kind of tempted to encourage making it a remote CAN device. Hmmm, so many things to consider.

Do my eyes deceive me, do I see prometheus uses 2 STM32's. It looks like the CJ125 is on the second chip, and that's not the same chip on our discovery board. Should we consider trying to get the analog on prometheus and frankenso to align? Or is that just crazy thinking. To me it appears that prometheus has a very different compile process, so I suspect that attempting to merge the designs doesn't help minimize software items. So I won't think about attempting to connect this board to the prometheus analog lines.

Re: CJ125 board

Posted: Tue Jan 30, 2018 3:55 am
by AndreyB
Yes remote CAN device would be great but who has the time to work on it right now? We are probably a few incremental steps between CAN device and where we are right now. Yes Frankenso add-on is not perfect solution. But for our next step we have two options
1) dozen wires, no hard connection between boards
2) some hard connection between boards and dozen minus 3 wires.

I am hoping that the extra effort to go from option #1 to option #2 is justified by the (small) benefit of having some hard connection and reducing the number of wires between boards.
kb1gtt wrote:
Tue Jan 30, 2018 3:42 am
Do my eyes deceive me, do I see Prometheus uses 2 STM32's.
Prometheus has a dual-chip design as in your can mount either a smaller more available chip and get limited functionality, or get a larger chip less available chip and get everything. I think Prometheus and Frankenso are too far apart to try to align anything, let the firmware flexibility worry about that.

Re: CJ125 board

Posted: Sat Feb 10, 2018 11:12 am
by AndreyB
Realized I have not yet placed the order, went to place the order. Do we want the whole cut-out thing at all? These mounting taps would be weak would they?

Also I wonder if the right tap would even be there - I think we have a board edge line where we should not.

Re: CJ125 board

Posted: Sat Feb 10, 2018 4:23 pm
by kb1gtt
Edge was mistaken, it should be corrected now, and recommitted.

I'm kind of 50% on if the mounting holes should be there. Those are mostly for Frankenstein folks, or other non-frankenso applications. I guess a small brain board could be made and this could be a kind of remote option. Any how, it has been updated.

Re: CJ125 board

Posted: Sun Feb 11, 2018 4:37 pm
by AndreyB
Just ordered some 0.3 boards

Re: CJ125 board

Posted: Thu Feb 22, 2018 1:40 pm
by AndreyB
@ has pointed us at https://www.etas.com/download-center-files/products_ES600/ES63x.1_UG_R04_EN.pdf ES630.1/ ES635.1 Lambda Module (1-CH) User’s Guide which has some interesting stuff about calibration.

"In this process, reference gas passes through the fitting. We recommend a concentration
of 8.29% O2 in N2 for the lean range. "
"To measure the sensor curve in the zero current range, a reference gas passes
through a washing bottle and the fitting. We recommend pure nitrogen at a
concentration of min. 99.999%."
"To measure the sensor curve in the rich range, a reference gas is pumped
through the fitting via a washing bottle. We recommend a gas mixture of 4.1%
CO and 3.2% H2 in N2. "

Re: CJ125 board

Posted: Thu Feb 22, 2018 2:52 pm
by Rhinoman
The CJ125 uses the calibration resistor in the lambda sensor's connector so we shouldn't need to do any calibration.