silverxxx's blog View Details
Posted by silverxxx | May 10, 2018 @ 06:49 PM | 16,075 Views
Roland2025 modified the acro code to work on his syma x5SC. It needs to be SC version, which has a stm32 processor ( the C version has a Nuvoton processor). I can't test this but I'm sure it's fine.

I'll probably ask him to merge this in the cg023 code when it's finished

Happy flying
Posted by silverxxx | Sep 25, 2017 @ 05:20 AM | 27,533 Views
Found this "Stunt" quad at kmart ( Australia ), for $20

Some info here:
protocol FQ777 ( devo ok nrf24)
tx- mode 2
battery - 150mA soldered
Weight 31g

Uses geared motors

This is a slowish quad intended for beginners, seems ok otherwise, not a lot of power, there is some more if the prop guards are removed. Battery door can also be removed. I intend to install acro in it to test geared motors.

Board inside uses a gd32 cpu, unknown radio and bosch gyro.
Posted by silverxxx | Sep 17, 2017 @ 10:59 PM | 27,366 Views
received an abovementioned transmitter today, it's the first commercially manufactured tx to ship with DeviationTx firmware ( opensource ). That's good, because it encourages diversity, even if it is of the software kind.

Here follows a quick review of this product:

The tx comes with a carry case, and a neck strap. The gimbals appear nice and smooth, and they have a similar size to the 7e gimbals, although very different mechanically. There is no feedback adjustment for them, except for throttle, which has a choice of smooth or clicky feel.

The lcd is slightly smaller compared to a 7e, it's 4.2x2 cm. Visibility appears ok, but I have not used it in sunlight.

The tx has 4 switches, all 3way. 4 trim switches also present. The 6 menu selection buttons are functional, there are 2 aux pots, these seem harder to use, they probably would be better with different knobs. I would not want to have to use the aux pots in the middle of a flight.

Mode conversion:

The tx arrived with Mode2, I needed to have it Mode 1

Sadly this was slightly harder then I thought, mainly after removing the centering spring, it turns out I also have to swap over the lever underneath, which required partial dissasambly of the gimbal. Turns out only 1 end has to be removed, that's enough to pull the level out.

First flight:

Strange, but once I set the tx to mode 1, the roll and yaw were inverted. I mean, ccw yaw makes cw turns. I believe this might be some firmware issue, but maybe I'm wrong. Anyway,...Continue Reading
Posted by silverxxx | Jun 07, 2017 @ 05:45 AM | 20,513 Views
Here is some info about the F08 cam + dvr , a soon to be released cam from Boldclash

F08-2 (4 min 55 sec)

It's a light cam, at about 5g, and as far as I know it's the lightest cam + vtx + dvr combo. The cam is slightly wider compared to a F01 cam. It has 6 bands and 8 channels each. Current consumption is about 600mA at 4.2V, voltage range is up to 5.5V ( not tested ) . The dvr has a "automatic save" feature, which means it splits the videos into 3 minutes each segment.

On the camera there are 2 buttons, one changes band / channel, the other starts/ stops recording

Edit: The B88 is now available, its $28 at this point on the boldclash site

A version with clover antenna is also available

Name: 20170606_113853_res.jpg
Views: 101
Size: 136.6 KB
Posted by silverxxx | May 03, 2017 @ 10:28 PM | 20,846 Views
So, I finally decided to use an openlog board to record some flight data.

First impressions: Although the openlog connects to an arduino minipro style ftdi, it has reversed serial connection for some reasons... Why would they do that? Configuration is done via a config.txt file, which gets written to the sd card if not present. So I'll change that file, it's the shortest way.

The 4 pins don't seem to be connected to a serial, maybe ISP? No info apparent on their function. I was hoping they are rx and tx, so I don't have to solder the ftdi, but they arent.

One interesting but likely nuisance feature is that the openlog will reset the config file if the rx pin is low at powerup...This can be disabled by setting the config option "ignoreRX" to 1.
Weight is about 1.6g without card
Name: 20170504_133040res.jpg
Views: 98
Size: 139.2 KB

I have the openlog flying and recording data now, wasn't too hard but there were some unusual things. So far it works ok off the 1s battery, I only got down to about 3V under load, but no issues apparent.
Name: 20170504_172046res.jpg
Views: 134
Size: 146.6 KB

Do and donts:
Apparently don't unplug the sd card while powered, definitely don't plug a new one in afterwards as it will corrupt it. Unplugging seems ok if the card is not written to at the time.

The board also flashes 3 times if there is no sd-card present, or it's formatted in a different way.
Posted by silverxxx | Apr 26, 2017 @ 05:08 AM | 33,559 Views
Keil compilation instructions for silverware:

Compiling a firmware with custom options is done using the software program Keil, on windows. This currently is the correct procedure for Bwhoop b03, E011 and also H8 mini with blue board. Use the right code for the board / quad please, they do not work correctly otherwise.

Software installation:

1) Install Keil MDK from
Note: right click the downloaded file and select "run as administrator", or there may not be enough folder access rights.

2) Install STM32F030 cpu support pack using the pack installer in Keil uVision. Keil uVision will ask to install the correct pack when the project file is open, if not installed. Choose "install", and wait for installation to complete (download). This will take some time.

Compilation process:

Assuming the installation process was performed correctly, we will now use Keil to compile a "binary", which is the file containing the firmware of the quadcopter.

1) download the project files from
Unzip to a folder of your choice.

2) open project file "/Silverware/silverware.uvprojx" or similar in the Keil uVision program. This can be done by double-clicking on the file for example.
Keil uVision may ask to install device support if not present, click "install", and wait for completion, which may take some time as it is downloaded.

3) Change the settings to your preference, generally using config.h and pid.c for acro pids.

3) compile the project and make sure there are zero errors reported in the lower part of the window. Use either the "build" toolbar icon, menu "project/build target" or "F7" hotkey. Keil will also save changes to files at his point.

4) Open the folder "/Silverware/Objects" ( name may change slightly) and find file "bwhoop.hex". This is the firmware file, to be flashed to the quadcopter using the "st-link utility" program.


Code location:

Keil download page:

Posted by silverxxx | Nov 09, 2016 @ 09:06 PM | 22,367 Views
Devo thoughts

My devo 7e arrived yesterday, and I flashed at the time, but I had nothing to fly with dsm protocol.

Today I did the module mod, I had only a normal nrf24 module around, so I used that. Ordering a module with the fast shipped order could have delayed the whole thing, so I did not do it. But it did get here in 1 week, for under a dollar expedited shipping. Good value? I think it is.

I liked the way the devo 7e is assembled, and it seems it had some good engineering put into it. I'm not a fan of products you take apart, and either a bunch of springs pop out, or lots of clips break for example. Not the case here.

After using the taranis gimbals for a long time, the smaller devo ones feel kind of like a downgrade, but I don't seem to have any problems, except it seems to me the devo gimbals have less precision very close to center. I think I might need to adjust the feedback springs a little.

So , now I have to figure out how to program a protocol for it, so I can have telemetry. I'll save that for tomorrow, I'm sure there is info on the net.

In conclusion, the devo 7e is very good bang for buck, imo it's probably no match for a taranis, but it's also 1/4 of the price. I think anyone who spent $200 or more combined on the rc hobby, and does not have one, or a better tx, should immediately purchase one.

The flash procedure was easy, and the nrf installation ok considering the device is easy to disassemble. Soldering not exactly beginner grade, though.
Posted by silverxxx | Aug 16, 2016 @ 05:13 PM | 73,136 Views
The h8 blue board has a strange issue when flashed with "Keil", and it needs specific workarounds.

Also, the board pads are mislabeled, the pad marked CLK is actually DAT, and viceversa.


This only affects the blue board and the Keil program (windows)

The actual flashing instructions are similar to those for the CG023 and other stm32 based quads, in this thread

An error comes up after the st-link is disconnected from the usb port . This results in a "communication error" popup.

To clear,

1. disconnect the st-link from usb and reconnect

2. Use the st-link utility (mirror) , connect to the board, and then disconnect, using the toolbar buttons or menu items

At this point the board should be flashable again in keil, until the usb is disconnected again.

Alternatives to the st-link utility:
You could also clear the error in other ways:
1: Flash any other non-blue quad using the st-link
2: Run the ST-LinkUpgrade.exe utility present in your keil folder, and upgrade the st-link
3: Use openocd to connect to a board
4: (included in above instructions) Use the st-link utility to connect to a board

Note: the board needs to be powered by a battery while flashing / erasing.

Sorry about this, the bug is beyond my control.

Hex file:
The hex file compiled by keil can be found in the "Objects" folder , h8blue.hex , and can be flashed with st-utility if needed

Use instructions in file . There should not be any issues except that the board pads are mislabelled. Gcc compiles somewhat larger, and as such you may need to turn off some features and watch file size, as the defaults are not always checked to fit compilation by gcc. You can see compilation status at travisCI

Code: (github)

Video of the procedure by gioy808 (rcgroups)

Video by ArbitraryNotion: (rcgroups)

Flashing using the st-utility only (post)

Binary location(use only if you have trouble compiling your own) (github)

Devo telemetry setup: RCG Post
Posted by silverxxx | May 22, 2016 @ 09:36 PM | 23,805 Views
I had a broken CG023 quadcopter board which had issues from the factory, I decided to use it as a plane rx, by flashing it with a stripped down version of the CG023 firmware

I was replacing a flysky receiver, so I got to save a about 3 grams. The board is approx 2g, which is not so bad, and rx range about 100m with a TX module using a nrf24 + pa board.

In using this rx board, I also intend to test several features:
*controls - normal mode
*controls + roll gyro
*controls + roll pid (p only)
*controls + roll pid (p and only)
*other stuff as it fits in place at the time
*maybe telemetry and some rx channel hopping algorithms

The plane I have it on now is a fairly low aspect ratio delta wing, which I built a while ago for other purposes.
Posted by silverxxx | Apr 22, 2016 @ 10:13 PM | 23,942 Views
When making a quad fly well, it is well known that a linear relationship is best. That is because pids work well with linear control systems.

So, a lot of methods have been devised to linearize the motors, and sometimes the effects of voltage sag and other effects.

Which leaves the following question: Yaw is created by propeller tip drag, so will linear thrust equal linear yaw aka tip drag?

it would seem such question would have been answered, but I have not found an easy answer in my short minutes of searches, so I decided to do my own measurements.

Name: yawvspwm.jpg
Views: 159
Size: 318.0 KB

As it can be seen , the thrust is already pretty linear. And the yaw is quite non-linear!

720 motors and hubsan 107 props were used. A pwm frequency of 8k resulted in almost linear thrust for this motor-prop combination.

An arduino based measurement system provided the results.

Another strange observation is how little (sideways) thrust the props produce compared to downwards thrust.

You can see the sub 2g forces are pretty noisy due to the measurement system limits. For the yaw test the motor was rotated 90 degrees.
Posted by silverxxx | Mar 01, 2016 @ 12:07 AM | 24,596 Views
I received a "1500mW 1.2Ghz transmitter" recently and proceeded to take it apart. Did not expect this tx to be actually 1.5W, but you never know. The price of this item was approx $43USD ( with receiver)

Name: 123.jpg
Views: 145
Size: 12.3 KB

Note: there are many similar looking transmitters, I do not imply that all of them are the same as this inside

Name: IMAG0465.jpg
Views: 260
Size: 524.4 KB
Description: 1500mW transmitter ( 1.2Ghz ) with case removed

I removed the case, which is the same weight as the tx innards itself. Without the case, it gets hot, but it might be ok even without airflow. Was ok for a few minutes.

...Continue Reading
Posted by silverxxx | Dec 14, 2015 @ 10:10 PM | 24,540 Views
I ordered and build a Hacker "mini reno racers" series plane recently, and as there does not seem to be any information about them I am posting my impressions here.

I bought the plane mainly because of it's low weight. It's advertised as 110g all included, which sounded great because it would be the lightest purchased plane I own.

The plane is designed and made very well, and getting the dihedral right was very easy, which was one of my concerns.

I do not usually build kit planes, as I am more likely to build my own, but I did notice some unusual design decisions made for this plane. For example the aileron servo is on the underside of the wing. Also there are no precut places for battery, esc and rx, or servos for that matter.

I actually think it's nice to be able to choose where to place the servo rather then to have the place ( and servo size ) chosen for you. However cutting holes for rx and especially battery can be a bit hard, since they are larger and the battery location may need to change.

Another different thing is the aluminium servo rod and the use of heatshrink for a hinge. I don't like to experiment with things like the servo linkages when normal steel wire and "L" horns work fine, so I went with a more standard way.

Name: IMAG0439res.jpg
Views: 138
Size: 356.5 KB

Name: IMAG0438res.jpg
Views: 146
Size: 1.02 MB

...Continue Reading