Re: [Paparazzi-devel] Umarim hardware question

[Karoly Molnar]

Hi Martin and Michel & Matthieu

Thanks for the answers.

On the 4017 issue:
I did not check earlier paparazzi autopilot designs, probably I ought have been to, before I blame Umarim. *Still* guys, please connect the 4017 VCC to 5V at least in the next designs. Then you would drive two short PCB lines from the MCU to the 4017 where the signal levels are not perfectly matching. This is still much better from any perspective than driving the potentially long servo lines with 3.3V. Certainly the best would be to use a proper voltage level translator.

On the barometer issue:
I am OK with the selection of the sensor and the concern on the lack of available component with I2C. On the ADC, well, you have made a choice that requires a number of companion chips with its additional cost, footprint and additional fault possibility. This is still OK, but now I understand that the driving factor in selecting components was the availability.
However using a voltage reference as power source is very hard to justify, especially that it is used beyond its normal operating conditions. I would rather skip the reference completely and power the pressure sensor from the 5V switchmode IC directly. On the second ADC channel I would measure the 5V thru a voltage divider and then the power supply variation could be immediately compensated in software.

On the switchmode power chip issue:
I could be wrong here. Looked strange with bare eyes and other QFN packages that I used have had smaller central pads.

I still respect what you have done guys, hope you are open to constructive criticism. And please, do not to drive the components beyond their normal operating range.

Regards
Karoly


> Message: 2
> Date: Mon, 5 Mar 2012 13:15:42 +0100 (CET)
> From: Martin Mueller <address@hidden>
> To: address@hidden
> Subject: Re: [Paparazzi-devel] Umarim hardware question
> Message-ID:
> <address@hidden>
> Content-Type: text/plain; charset="utf-8"
>
> Hi Karoly,
>
> not sure about the ITG3200/power switch footprints but for the 3.3V-only
> pulse signal from the 4017 to the servos: It has been like that for all
> autopilot boards since the first Tiny and I am not not aware of any issues
> related to that.
>
> Martin
>
>
>
>
> Karoly Molnar <address@hidden> hat am 5. M?rz 2012 um 11:47
> geschrieben:
>
> >
> > Hi All,
> >
> > Appears that no one could answer my questions in my first mail. In the
> meantime I have found a number of other design flaws with the Umarim
> hardware:
> >
> > 1. Apart from the ITG3200 solder pad issue that is discussed in a
> separate thread, there is another pad size issue with the switchmode power
> supply IC, even if this is not that serious. The gap between the center pad
> and the signal pads is just too small. I had short-circuit here with one of
> the boards that I tried to solder. Still it is not a serious issue and easy
> to fix in the gerber file directly (actually you do not need Protel to edit
> a gerber file).
> >
> > 2. The 4017 IC that distributes the servo signals is powered from 3.3V
> which might be adequate to drive the 5V servos but certainly not in the
> guaranteed limits and could be a source of noise related issues if longer
> servo wires are used (most servos have CMOS input not TTL). In case the
> servos are driven from 6V which is also common in certain applications, the
> 3.3V HI level is just unacceptable.
> >
> > From the list of hardware issues I shall conclude that Umarim needs a
> V2.0 review; the 1.0 is just not reliable enough, even if it is a nice try
> in this small form factor. I will certainly not finish the manufacturing of
> the 3 PCBs that I have purchased but look for another hardware solution.
> >
> > Regards
> > Karoly Molnar
> >
> > From: address@hidden
> > To: address@hidden
> > Subject: Umarim hardware question
> > Date: Thu, 1 Mar 2012 20:32:43 +0100
> >
> >
> >
> >
> >
> >
> >
> > Dear All,
> > I have a number of questions to the Umarim hardware design. I really
> appreciate the effort that someone has put in designing and testing this
> circuit so please do not interpret my questions offensive. There might be a
> certain motivation behind the design decisions that I do not see yet.
> > 1. The pressure sensor is powered from a voltage reference. The
> guaranteed output current of this reference is 5mA while the average
> current consumption of the pressure sensor is 6mA. I guess the designer
> wanted to keep the voltage as precise as possible in order to have a stable
> output value but the reference circuit is apparently operating beyond its
> normal working conditions so this goal is hardly achieved.2. Also for the
> pressure sensor, I see here a 16bit ADC connected to it that is powered
> from 5V and then connected an I2C level translator. Couldn't have been
> simpler to power the ADC from 3.3 V and match the voltages on the ADC
> input?3. Also there are I2C pressure sensors available, why wasn't that
> used?
> > I am just guessing that the complicated design has the aim to provide an
> extremely precious pressure reference. I am still beginner in autopilots so
> it could be the case that such accuracy is really needed, but at first
> glance it seems just too complicated to use 5 integrated circuits to get a
> pressure value.
> > What do you think?
> >
> > RegardsKaroly Molnar
> >
> >
> >
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://lists.nongnu.org/archive/html/paparazzi-devel/attachments/20120305/185aae7b/attachment.html>
>
> ------------------------------
>
> Message: 3
> Date: Mon, 5 Mar 2012 13:44:54 +0100
> From: Michel GORRAZ <address@hidden>
> To: address@hidden
> Subject: Re: [Paparazzi-devel] Umarim hardware question
> Message-ID:
> <address@hidden>
> Content-Type: text/plain; charset="iso-8859-1"
>
> Hello,
>
> 1. Ok, there is footprint issue on the ITG3200 Gyroscope. Sorry for that.
> We did'nt notice this problem on all Umarims we assembled by ourselves.
> 2.* "...The LT6656 can supply up to 5mA of output drive with 65ppm/mA of
> load regulation...".* Load regulation should be a bit worse at 6mA (output
> voltage change should be around -250ppm, see "*Load Regulation / output
> voltage change*")
> 3. ADS1114 internal reference voltage is 4.096v, so it must be 5v powered.
> ADS1114 was (for us) the best choice at this time : 16bits/I2C/fast and
> small enough, easily available.
> 4. when we started the design of the Umarim, there wasn't lots of
> choicesin good sensitivity (and easily available) digital I2C
> barometers.
> 5. We don't use 5 ICs instead of one just for fun...
> 6. TPS62112 switching power supply IC footprint is perfectly matching the
> recommended land pattern in datasheet.
> 7. For the 4017, see Martin's perfect answer.
>
> Except for Barometer design, declaring Umarim not reliable seems to be a
> bit...excessive.
>
> You guys, what do you think about an Umarim "Lite" without barometer ?
> (while waiting v2.0 Umarim)
>
> @+
> Michel & Matthieu
>
>
>
>
> 2012/3/5 Karoly Molnar <address@hidden>
>
> > Hi All,
> >
> > Appears that no one could answer my questions in my first mail. In the
> > meantime I have found a number of other design flaws with the Umarim
> > hardware:
> >
> > 1. Apart from the ITG3200 solder pad issue that is discussed in a separate
> > thread, there is another pad size issue with the switchmode power supply
> > IC, even if this is not that serious. The gap between the center pad and
> > the signal pads is just too small. I had short-circuit here with one of the
> > boards that I tried to solder. Still it is not a serious issue and easy to
> > fix in the gerber file directly (actually you do not need Protel to edit a
> > gerber file).
> >
> > 2. The 4017 IC that distributes the servo signals is powered from 3.3V
> > which might be adequate to drive the 5V servos but certainly not in the
> > guaranteed limits and could be a source of noise related issues if longer
> > servo wires are used (most servos have CMOS input not TTL). In case the
> > servos are driven from 6V which is also common in certain applications, the
> > 3.3V HI level is just unacceptable.
> >
> > From the list of hardware issues I shall conclude that Umarim needs a V2.0
> > review; the 1.0 is just not reliable enough, even if it is a nice try in
> > this small form factor. I will certainly not finish the manufacturing of
> > the 3 PCBs that I have purchased but look for another hardware solution.
> >
> > Regards
> > Karoly Molnar
> >
> > ------------------------------
> > From: address@hidden
> > To: address@hidden
> > Subject: Umarim hardware question
> > Date: Thu, 1 Mar 2012 20:32:43 +0100
> >
> >
> > Dear All,
> >
> > I have a number of questions to the Umarim hardware design. I really
> > appreciate the effort that someone has put in designing and testing this
> > circuit so please do not interpret my questions offensive. There might be a
> > certain motivation behind the design decisions that I do not see yet.
> >
> > 1. The pressure sensor is powered from a voltage reference. The guaranteed
> > output current of this reference is 5mA while the average current
> > consumption of the pressure sensor is 6mA. I guess the designer wanted to
> > keep the voltage as precise as possible in order to have a stable output
> > value but the reference circuit is apparently operating beyond its normal
> > working conditions so this goal is hardly achieved.
> > 2. Also for the pressure sensor, I see here a 16bit ADC connected to it
> > that is powered from 5V and then connected an I2C level translator.
> > Couldn't have been simpler to power the ADC from 3.3 V and match the
> > voltages on the ADC input?
> > 3. Also there are I2C pressure sensors available, why wasn't that used?
> >
> > I am just guessing that the complicated design has the aim to provide an
> > extremely precious pressure reference. I am still beginner in autopilots so
> > it could be the case that such accuracy is really needed, but at first
> > glance it seems just too complicated to use 5 integrated circuits to get a
> > pressure value.
> >
> > What do you think?
> >
> >
> > Regards
> > Karoly Molnar
> >
> >
> >
> >
> > _______________________________________________
> > Paparazzi-devel mailing list
> > address@hidden
> > https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
> >
> >
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://lists.nongnu.org/archive/html/paparazzi-devel/attachments/20120305/fa06b5c5/attachment.html>
>
> ------------------------------
>
> _______________________________________________
> Paparazzi-devel mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>
>
> End of Paparazzi-devel Digest, Vol 96, Issue 15
> ***********************************************