DIYMore Pro-Mini board issue

[zep]

I have made several Kraft TX conversions to 2.4GHz based on Phil’s 7-channel propo design using the DIYMore Pro-Mini encoder board and they all work quite well. I recently bought two more DIYMore boards and found they both have what appears to be noisy A/D circuit in the 328P chip.
The effect is that at max stick deflection and max Expo the widths of the pulses are unstable resulting in significant jittery servo arm movement. The unstable pulse widths are clearly visible on oscilloscope. The jitter is less with lower Expo setting and/or reduced stick deflection. Both new boards have the same amount of jitter. My older Kraft transmitters with same wiring and code have no jitter at all.

So I replaced the new boards with one from an older TX and found no jitter. Then I substituted the stick pots with pots on very short leads and the jitter in new boards mostly subsided. Per Phil’s suggestion I put a 33uF (he suggested 100uF) tantalum cap across +5V to GND. This cured the problem and now there is only the slightest of jitter with max stick deflection and max Expo. No observable servo arm movement, just feel a very slight twitching when I hold the servo in hand.

So my question is, have any fellow nerds here observed this problem with the DIYMore Pro-Mini board?

Thanks,
Tommy

[Pchristy]

Not had the problem with the DIYMore board, but I did have a similar problem on my Ace MicroPro 8000 conversion. (Its a very early computer transmitter, dating from the early 90s, but surprisingly sophisticated for its age!).

Two things cured it in my case. One was replacing the battery pack with a newer (lower internal resistance?) one, and more importantly, clipping ferrite suppressors around the leads connecting the RF board to the encoder.

Your mileage may differ, but it might be worth a try.

--
Pete

[zep]

Thanks for the input. I observe the noise issue with my two new boards without the RF module, receiver or servos in the path. The instability is evident on scope at the PPM output with nothing else connected. Also with my version of Phil's TX PPM analyzer.

[Tobe]

Could be worth trying twinning the wires to the RF board

[zep]

Thank you. The problem exists with or without RF module connected.

[Mike_K]

There are so many things that could be causing this, it will be a case of eliminating each possible cause, one by one. Is it all channels affected or only one channel?

If it is just one channel then a noisy pot is probably the cause. I guess it is possible that multiple pots are noisy if the transmitter has been stored in a damp place or the battery has leaked. Disconnect all the pots and try just a couple of test pots and link all other analogue inputs to 0V or 5V and test it to check it isn’t the pots. Even two fixed resistors could be used in the place of a pot, if you haven’t any spare pots, you are trying to get a stable ppm output not checking the joystick outputs.

If all channels and the sync pulse timings are affected, then there could be something wrong with the voltage regulator or something is wrong with the joystick pot wiring.

Try powering it temporarily with a 5V uBEC, the 5V BEC on an ESC or a 4S NiMH battery (not fresh off the charger when it is nearer 6V, but discharged to nearer 5V), just plug it into a spare input D8 to D13. Also, ensure any unused analogue input is linked to ground or 5V (a header jumper or Spektrum style bind plug can be used). If it works OK with the external 5V supply, you have a faulty voltage regulator or faulty on-board decoupling capacitors. Try a 10uF on 5V and a 100nF from Aref to 0V. If that doesn’t work it’s time to get another Pro Mini Strong from a different supplier (and a different batch).

It is worth testing the resistance across the 5V output (powered off), you should see the resistance of all the pots in parallel, if you have say 5 pots and each is 5K, you should see approximately 1K resistance. If it is way lower, try unplugging one pot at a time to find the faulty one.

[zep]

Thanks Mike.

All 4 stick channels are noisy with max deflection and max expo. Slight jitter at neutral. The aux channel has pot (no expo) and it is stable. I substituted stick pots with pots on very short leads and the jitter is gone. All 8 analog inputs are loaded. Seven pots and one channel has 4-way Trim switch which has 20k pulldown resistor in open position. I replaced the Trim switch with jumper to GND and still jitter. I tried with different 2s lipos and a switching power supply from 7.0V to 8.4V and still jitter.

I put 10uF cap on 5V and no jitter. I plugged 5V BEC into 5V bus without the 10uF and no jitter.
I thought I had looked at 5V regulator noise before but apparently not close enough. Now I see 400mv p-p noise, about 20kHz oscillation. The cap kills it, <5mv p-p noise and no jitter.

Without the cap there is slight jitter at full stick deflection and expo pot set in the middle. So it seems the regulator noise is banging the pulsewidth and translates to significant jitter on the steep slope of expo curve.

[Mike_K]

if you have 400mV ripple @ 20KHz, no wonder the ADC is having a hard time getting a stable reading. The ADC charge and hold circuit will have a "different" 5V voltage to compare to each time... I would guess it is caused by faulty decoupling caps. Or have they even been fitted? It is worth checking that they are physically there as it would not be the first time that the "pick and place" machine has failed and not fitted the smd components. You can also get voltage regulator output ripple if ultra-low ESR caps have been used, but as these caps are expensive and DM Moore boards are budget it is unlikely.

You now have to decide if to use the Pro Mini Strong with external smoothing caps or get a new Arduino. If you do decide to use it, do fit a 100nF cap between the Aref and 0V, it will help the ADC have stable measurements.

[Phil_G]

In a weird sort of way this is both reassuring and amusing, we'd been discussing Tommy's problem by PM on RCG and the diagnosis went exactly as per the posts above, I'd suggested Tommy also ask here but deliberately kept out of it to see what suggestions he got
We know that great minds think alike, but numptys like me can think alike too!
I would suspect the regulator output cap but if you fancy changing the SMD regulator too, it comes off quite easily Tommy - I've had some 16Mhz ones with 3v3 regs, their quality control isnt what it used to be, although the grounded A6 or A7 (or both) problem seems to have gone away, my last batch were fine

[zep]

I haven't found missing caps or wrong regulators on my boards. I replaced the larger cap near DTR pin with 10uF Kemet T491B106M016AT surface mount on hand. A smaller case size would fit better. Noise on 5V now <5mv p-p. So that's my solution for now. In the past I have had blown regulators, maybe ESD or my wiring error. I replaced them with TLV1117-50CDCY and seem to work OK.

I should have figured out this one myself but it has been a fun trip nevertheless. And I appreciate all the input and troubleshooting ideas, good brain therapy.

Cheers,
Tommy