Another way to hook up a DSM module.

[akschu]

Looking for some thoughts and feedback on how I did this. The ideas is to get the best reliability possible since I plan to fly expensive stuff with my dsm2 conversion.

I took a DSM2/DSMX module out of a DX4e then used this device to do the ttl level conversion:

https://www.sparkfun.com/products/10403

Here is the datasheet:

http://dlnmh9ip6v2uc.cloudfront.net/dat ... CA9306.pdf

So on the input side I put the ppm pin from the tx on SDA1, then I found a 5v power tap from the first pin on the daughterboard as the VREF1 signal reference (white wire). On the output side I went from SDA2 to the DSM module (yellow wire), then I powered the module from a LF33ABV voltage regulator that is fed from the battery (through a switch) (red wire) and also used the output from the regulator as VREF2.

I have attached the schematic and the pictures. I *think* this is a better solution than using a diode to drop the voltage since this device is designed to allow bi-directional serial between devices with different voltages.

So far everything works fine. I have been flying little micro helicopters, but I want more feedback before I fly anything that can hurt someone.

So my questions:

1. Does this look sane to you guys?

2. How can I test to make sure that it's fine? I have a logic analyzer as well as a 2-channel o-scope. Should I check to see if the input and output signals match?

3. Any thoughts on my 5v tap? I'm just using it as a reference, but didn't take the time to trace it all out. I think I need to look at this on the scope and make sure it's a perfectly stable 5v. Perhaps trace it to the 5v regulator on board.

4. Do you think that this 5v regulator on the main board is used to create the ppm reference? If so then this should be fine, otherwise, it may have hidden issues.

thanks,
schu

[Kilrah]

Hmm, that looks like a nice component

I see one issue with your schematic when compared with the datasheet, you have your high side (5V) on the "left" (Vref1, SCL1, SDA1) and the low side on the Vref2 side. The datasheet mentions that:

- The Vref2 side is supposed to be the high side, from the application diagram on page 9:

The applied voltages at Vref(1) and Vpu(D) should be such that Vbias(ref)(2) is at least 1 V higher than
Vref(1) for best translator operation.

- Vref1 side should not be higher than 4.4V, Table 9 on page 10:

Vref(1) reference voltage (1) (Min) 0V (Typ) 1.5V (Max) 4.4V

And the general specs on page 1 say:

The PCA9306 is a dual bidirectional I2C-bus and SMBus voltage-level translator with an
enable (EN) input, and is operational from 1.0 V to 3.6 V (Vref(1)) and 1.8 V to 5.5 V
(Vbias(ref)(2)).

So, to do it right I'd swap the sides if I wanted best reliability. You likely haven't damaged the chip as the 5V on Vref1 still falls within the "Recommended operating conditions" on page 6.

Once that's done, I'd check the signal levels are correct on both sides (both high and low states) and not "forced", and that the signal is clean. I'd also check the 3.3V supply ripple on the DSM module. As it draws current very irregularly with serious spikes some regulators have trouble adapting quickly enough.

EDIT:
I've looked at your pics again and it seems that you actually wired it correctly, not according to your schematic.

[akschu]

Yea, nice catch. I did wire it correctly, but my schematic is wrong. I'll fix it and repost since I don't want someone following in my footsteps and running into trouble.

I'll break out my scope later today and test this stuff to make sure it's clean.

[rperkins]

I *think* this is a better solution than using a diode to drop the voltage since this device is designed to allow bi-directional serial between devices with different voltages.

Yes, but isnt the communications only one way ? From the microcontroller to the TX board. Your solution is 'more correct' , but IMHO I cant help but think a simple diode is sufficient in this case considering the cost and having to physically locate the small board somewhere.

[Kilrah]

I think he is concerned about the low level sent to the DSM module being the diode's dropout voltage (about 0.6V) instead of a "clean" 0V.

I've personally been flying with a diode AND the inline 1k resistor mod (throwing the low level even way higher, to about 1.1V) for many hours without a glitch, so there's a lot of margin... I've also never read about a problem with the diode version, so I'd say it's really a non-issue, but each to his own

[rperkins]

Thanks for the reply. I was thinking that I was overlooking , or missing, some aspect of it. That happens a lot

[ShowMaster]

I just opened up a DX4e I bought from a hobby shop that had several in a box along with other TX's they were getting rid of. The DSM module inside has DX5e on it so I hope this is the high power one?
SM

[calla969]

I'm pretty sure the DX4 and DX5 use the same module.

[cre8tiveleo]

That's what I keep saying, I'm not blowin smoke out my bum, They are the same.

[ShowMaster]

Ok, I'm counting on your Bum and installing it after I verify it's working.
Thank you and your Bum.


ShowMaster
Sent from my iPod touch using Tapatalk

[jhsa]

[ShowMaster]

I read a post about a recall on the dx4e dsmx losing dsm2 link. Since mine hasn't been robbed of parts yet I'm going to look into it. Anyone deal with this yet?


ShowMaster
Sent from my iPod touch using Tapatalk

[ShowMaster]

My PID sticker reads HO202F. According to the recall bulletin my tx is not on the recall list so I guess I'm good to go?
I may call Spektrum before I tear it apart anyway.


ShowMaster
Sent from my iPod touch using Tapatalk

[akschu]

The deal with the recall is that the firmware on the first DSMX rf modules caused the low power specktrum DSM2 modules (anything below a AR7000) to hang, then the watchdog resets them. So a good way to figure out if you are affected is to get the TX bound to an old AR6200 or AR6100 then leave it running for a few hours. You can move the sticks around and move the TX around, just lots of diversity, and eventually the RX will start blinking as if it had a brown out. It didn't actually brown out, but it did reset due to the watchdog kicking it after it hung.

The DX4e came out long after this bug was found, so I doubt you have this issue.

As far as my module goes, I did some looking on my scope last night. The 5v line I used is very stable. Very little ripple. The caps and voltage reg I used are doing fine as the 3.3v line has little ripple even though the spektrum module is powered on.

The signal lines look great. It goes from 0v to 3.3v in perfect sync with the ppm line from the TX. In fact, the signal is even cleaner because the signal on the TX goes to -40mv on the low side.

At this point I'm ready to start testing range. I'm going to do some ground based testing, and if I can get a good range check and confirm I can get 1000ft away on full power on the ground then I'll start flying my son's trainer with it. I'll also fly some of my smaller helis with it. Once I'm comfortable with it, I'll move up to 600 and 700 size helis.

[Kilrah]

Interestingly when I got my (2nd hand) DX4e I intended to scavenge the module from, I first bound it to my eflite umx plane and did a couple of flights. Not many because it wasn't fun to fly without expo, but enough to find that there was a recurring problem, when sometimes throttle would stop responding. Motor would just stop, I could then move the throttle stick up and down a few times without any effect, and it would jsut start working again after a couple of seconds. It did it with 2 sets of batteries, on different days, so it wasn't a one-time glitch. Other controls were still responding normally.

I took the TX apart, installed the module on my 9x, and never had a problem again. So apparently my DX4e's mainboard had some kind of issue...

[rperkins]

The DX4e came out long after this bug was found, so I doubt you have this issue.

I only fly an mcpx, which isnt dsmx. Dont know much about the issue. Just remember that there was a recall issue with the DX4e, as I was getting some off ebay for donor modules. The recall listed here may even be a completely different issue than you are discussing.

http://www.spektrumrc.com/Products/Supp ... d=SPMR4400


update - I have 3 dx4e's and none are affected. My pid's start with H0. One unit came in a RTF mcpx from last christmas, the other 2 were bought from online sources this spring.

ps .. Nice scope picture

[ShowMaster]

The DX4e came out long after this bug was found, so I doubt you have this issue.

I only fly an mcpx, which isnt dsmx. Dont know much about the issue. Just remember that there was a recall issue with the DX4e, as I was getting some off ebay for donor modules. The recall listed here may even be a completely different issue than you are discussing.

http://www.spektrumrc.com/Products/Supp ... d=SPMR4400


update - I have 3 dx4e's and none are affected. My pid's start with H0. One unit came in a RTF mcpx from last christmas, the other 2 were bought from online sources this spring.

ps .. Nice scope picture

Great news, thanks.
I waited on the Spektrum service phone line and my ear went numb so I left a call back number but after an hour I have to leave so I'm going to go with your HO series PID and mine and go with the fact that it's not listed as a recall series and yours works. Thanks again.

update
Got a call back after my post from a very helpful person and all my questions were answered and I'm OK.