Lately I had some problems with resetting the UHU Servo Controller Chip, but that is history.
In the meantime, something else occurred. After applying the 12 Volt power on a UHU Servo board, "POWER" led turns on, and after a little while "OK" led turns on. That is how it should be. But, yesterday, when I connected a board, which was working fine two weeks ago (like I mentioned in a sentence before), to a 12 Volt power supply, only "POWER" led turned on. I connected it to a computer, but no message "....UHU Servo Board.... EEPROM OK....." appeared, and you can not see or change PID parameters. So, while I was reading Uli Uber user manual, after a while, I looked on the board again, and it was fine!! Both "POWER" and "OK" led were on.
I suspect the next. The board was in a room that is not heated - only a few degrees (not more then 10 Celsius). I took it to a room in which I work (20 degrees Celsius), and as I said, after some time it start working. Don't know but I suspect that chips are temperature sensitive!
If somebody there had the same problems, be free to comment.
Last edited by mibo1112; 02-05-2010 at 04:15 AM.
And thanks for a replay.
I purchased boards in USA, from embeddedtronics. They look pretty cool, I mean the production of a board is really professional. So, I don't think that it is a board problem.
On my opinion, the problem are chips (but still didn't find which one), and I think that they are temperature sensitive. Though, I didn't put the boards outside on -10 Celsius or something! They were on maybe 7 or 8 degrees Celsius.
On a board there are:
1. 4013 RS Flip Flop - made by ST.
2. 74HC14 Schmitt trigger - ST.
3. MAX232 - don't remember a manufacturer, though it's only for a connection with a PC, so it shouldn't be the reason I think
4. IR2184 output transistor drivers - made by international rectifiers
5. TL 082.
You need to check all the soldering on that board. The power-LED is not driven from or controlled by the UHU-chip so the fact that IT doesn't come on has nothing to do with UHU chip itself. (If the chip itself isn't shorting out the powersupply of course).
If the OK-LED, which IS controlled by the chip does indeed come on it means that the UHU-chip is execting code. I suspect you either have a cold joint, bad IC socket, a short or something - somewhere. OR that there's a problem with the X-TAL on the board and the temperature makes it oscillate at the wrong frequency. This would still make the UHU-chip execute code but the baudrate for communicating with the PC would be wrong.
Well, that's two ideas....the ICs themselves should have NO problems at all operating at 10°C.
Sorry about that first post massage, I made a mistake, it's changed now. So, after powering up the board, only "POWER" led turns on (and not "...only the "OK" led..." - that was a mistake).
Ofcourse that there is a large posibillaty that "POWER" led will turn on, cause it is not driven by the UHU chip, and that the most probably reason for not turning on this led after powering up the board will be in board connections. So, not to make this story larger, I have no problem with the "POWER" led. Problem is with "OK" led.
Cold joint, and bad IC sockets, did cross my mind. I even made all soldering again on one board, but that didn't make the problem go away. Interesting thing is that I have two boards assambled, and I leave both of them, side by side, on the same spot, and after a day or two, sometimes more, when I take them and connect to a 12V power, the same thing occures ("POWER" led turns on, "OK" led not) on each board! And, after a while, the "OK" led turns on too.
Anyway, thanks for those ideas.
If You would be so kind to elaborate about that sentance of Yours "...OR that there's a problem with the X-TAL on the board and the temperature makes it oscillate at the wrong frequency...". I suppose that You ment about that 24MHz cristal oscillator putted on the UHU Chip.
Thanks for the ideas,
Yes, I meant the crystal oscillator. Did you get the crystals from Uli together with the UHU chips or have you bought them elsewhere? Apparently it has to be a 24Mhz, 1st order (whatever that means).
As long as the oscillator oscillates at ANY frequency the chip will execute the code but all timing will be off as the program "thinks" it's running at 24Mhz. And since the UART (that handles the communications with the PC) derives its clock from the main oscillator the baudrate will be off.
Well, like I said it's just a theory....
Do you have access to an oscillosope?
Yes, crystals were bought from Uli, cause there was a story that not all crystals will work. So, I didn't want to think about it and ordered them together with UHU Servo Controller chips.
Yes, I own an oscilloscope - TEKTRONIX 2215. It's a little bit old but works fine and serves the needs.
I didn't understand very well that thing You mentioned "...but all timing will be off as the program "thinks" it's running at 24Mhz. And since the UART (that handles the communications with the PC) derives its clock from the main oscillator the baudrate will be off...." What timing will be off? What program "thinks"? Mach or UHU Controller firmware or how to call it? What UART stands for?
Did You mean just to measure natural crystal frequency with oscilloscope?
I mean that IF the oscillator, for whatever reason (like wrong value on the caps) oscillates, but at the wrong frequency the chip will try to talk to the PC at a different baudrate which obviously won't work. The fact that the OK LED does turn on indicates that the chip IS running but the fact that it won't coomunicate with PC (which it did earlier I suppose) indicates that the frequency may be wrong.
Measuring on the oscillator can be tricky and might cause it to stop (if it IS running) but you can TRY and see what you get. Otherwise, when the OK LED turns ON and if the Error LED is OFF try measuring the PWM frequency, if it isn't ~20kHz then there's something wrong the oscillator.
The operating temperature of the UHU chip itself is -55 to 125°C so 10° shouldn't be a problem.
Can you measure the voltage at the UHU terminals, it should be 5V in theory. What power/voltage rating of PS you use for the logic supply?
Maybe you could try to remove the other IC's and see what happens. I've had a different problem with a bad batch of 7414's, so don't trust ICs.
I am not so sure, but I have read here on the forum that this particular kind of boards have some noise problems and chip resets. I cannot take a side as I have not used them. Mine is the first PCB, made 2005.
Henrik, just to be clear, I never had a problem that "OK" led illuminates and no connection with the PC is made. So, you take the board (relatively cold), connect it to a 12V power supply, and to a PC. Just those two connections. You've already started serv200.exe and going to turn the power on. When the power is on, only "POWER" led turns on. You have no message on the PC computer screen. Actually, nothing changes. Then you wait a little bit (half a minute or so) and the "OK" led turns on. The moment the "OK" led turns on, the message "...UHU Servo Controller ver. 3.00 ..." came up.
Now, I have 5 UHU Servo controller boards assembled. All of them are on the same spot (place). I'll try them all today, to see if all will have the same problem.
I measured crystal frequency yesterday. Tektronix is a 60 MHz, so it could be done. I measured on pin4 and on pin5 relatively to "GND" point of the controller board. Both times you see the "sine wave" on the scope.
I didn’t measure on the cold board, I'll do it today.
Couldn’t write earlier, cause of some "net" problems.
Anyway, one of two problematic boards I, let's just say resoldered (didn't have time to do that on both boards). After soldering (well, actually just heating again) all the contacts, board worked fine. Leaved her a day or two on a cold place, then tried it again, and both led turned on immediately - well, "OK" led just a little bit latter then "POWER" led, cause of that UHU chip reset procedure on start, but it's no mistake.
The bottom line, it was a connection problem somewhere. Well, to be honest, it smelled that way all the time, right?
Yes it did. I'm glad you found it though.