HelloChirp software | TDK InvenSense

By highvoltagecon… , 16 April 2020

Forums

Dear all,

I have question to HelloChirp example firmware. I have run short distance range configuration with max range 75 mm. My object is detected stable only when I am moving it (so called none-stationary object). When object stop there is a couple of good measurements and then series of messages "target not detected" in between. Is this related to the treshold detection levels? I found little note in datasheet saying it can be issue for 40 mm distances for none-stationary objects but i can observed instability for the range between 40-70 for max range of 75 mm and 40-85 mm for max range of 150 mm.

Do You have solution for that problem?. Is that software configuration issue or sensor itself can behave like this (it is crucial for my application to have range stable between 40-100mm for stationary objects).

I just installed the software and tried to use DK-CH101. Initially, I had the same problem as you had described. However, I tried other configurations and everything starts operating as expected. Try the attached configuration that worked for me.

phpbb Post ID

38331

The web application is not allowing the uploading if a png image..

phpbb Post ID

38333

But that's the config:
Sensor 0: TX_RX
Sensor Mode: Normal
FSR(mm): 1000
Sampling Rate: 10
STR: checked
A-Scan: non-checked

phpbb Post ID

38335

Hi, I have checked Your setup for ch101 and it looks like it does not work at all for short distances below 75mm (picture with xxx_normal_config1). For short distance and different FSR and SR it look like it is showing sometimes good result (49-51mm) but mostly wrong ones (see pictures with xxx_short_config_yy) any doubts what I am doing wrong?

phpbb Post ID

38341

it look like the pictures upload is very limited, i am applying log files which does not have security or size problems, from those 3 logs it is clear Your config does not see 46-51mm objects (stationary or none-stationary does not matter), short distance config sometimes see not moving objects but mostly not (log file does not show this since omitting 0s), short distance see perfectly short distances when objects moving (slightly shaking)

phpbb Post ID

38349

it look like cvs are not allowed as well so raw data I am sending for normal (Your config)
8528,110,25,6126,
8529,112,06,7741,
8530,161,19,13825,
8531,164,09,14015,
8532,106,03,4927,
8533,103,31,4749,
8534,104,28,5498,
8535,163,53,15507,
8536,163,69,14647,
8537,104,41,7556,
8538,122,44,4703,
8539,164,16302,
8540,144,22,14081,
8541,163,47,16554,
8542,108,53,3846,
8543,164,28,15759,
8544,164,41,16730,
8545,105,75,4158,
8546,164,06,17352,
8547,149,19,16571,
8548,164,31,16459,
8549,108,88,5278,
8550,163,72,17500,
8551,104,03,4880,
8552,107,03,5471,
8553,114,12,5649,

short not moving obj (mostly 0s sometimes good point):
279,0,3311,
280,0,3311,
281,0,3311,
282,0,3311,
283,0,3311,
284,0,3311,
285,0,3311,
286,0,3311,
287,0,3311,
288,47,84,3057,
289,0,3057,
290,0,3057,
291,0,3057,
292,0,3057,
293,0,3057,
294,0,3057,
295,0,3057,
296,0,3057,
297,0,3057,
298,0,3057,
299,0,3057,
300,0,3057,
301,0,3057,
302,0,3057,
303,0,3057,
304,0,3057,
305,0,3057,
306,0,3057,
307,0,3057,
308,0,3057,
309,0,3057,
310,0,3057,

short moving object (shaking): mostly good points
675,47,06,9597,
676,46,5,9084,
677,47,25,10153,
678,46,44,9864,
679,46,44,9616,
680,45,31,8398,
681,45,44,8444,
682,45,47,8036,
683,45,53,8131,
684,46,16,9657,
685,45,31,7686,
686,45,34,7571,
687,46,06,9247,
688,46,28,9417,
689,45,62,7693,
690,46,16,10865,
691,45,34,7655,
692,46,12,8147,
693,45,69,7620,
694,45,28,11593,
695,45,41,11259,
696,45,62,7892,
697,45,28,9450,
698,45,69,10442,
699,45,84,10102,
700,45,19,9705,
701,45,56,9290,

phpbb Post ID

38350

I don't have much experience with the DK-CH101. The majority of my tests were between 2mm..40mm under that configuration. It worked pretty nice since it was able even to capture the light shaking of the hand (I was handing a paper box as obstacle). However, I remember that I moved quickly to a higher position and the response was 0 all time and few seconds later, the correct distance. The company failed in providing a good block diagram for the CH101 sensor and its Evaluation Kit.

It seems that CH101 has a microcontroller or DSP inside since it has the I2C interface. The ts3a27518e chip after them is a 6-channel MUX - I had not investigated the details, but it may activate one sensor at each slot time and direct the I2C channel of that selected CH101 sensor to the main MCU at the evaluation board. Therefore, the characteristics of the 0...0 response issue may be caused by software bug at the Evaluation Board. I would like to investigate deeper but I don't know if I will have time.

One line of research would be to use their simple "Hello chirp" example. This may help to clarify if the main MCU must send any kind of special signal to CH101. I don't know why they mention CHIRP all the time. If the CH101 is an autonomous unit, all the chirp generation and time of flight calculations would be done inside CH101 and we only have to request and get the distance using I2C and we don't have to care about low-level details as chirps. However, if CH101 is not autonomous, that's a problem for me since there are many other options at the market that are simple and "ready-to-use" without having to care about timing calculations. Again, the InvenSense company really failed to show these details at the 1-page datasheet they provided.

As a last note, while looking for CH201 information, I found this document: https://www.tdk-electronics.tdk.com/download/2431644/f7219af118484fa9afc46dc1699bacca/02-presentation-summary.pdf

It has many of the answers for the above questions, in particular the fact that CH101 is autonomous (full SoC). In this case, give a try and try to connect the sensor directly to your microcontroller - it seems that the sensor is good, but the bug is at the evaluation board side (hw or fw).

phpbb Post ID

38351

Hi,

It is amazing what you said to me that You have achieved stable stationary object detection between 2-40mm range (they said in datasheet minimum range is 40mm and forty only for none stationary objects)? Did You used DK-CH101 board or Your custome board? It looks like Your config is used for normal range detection of 1m. They said that if we want to detect shorter distances we need short distance mode not normal. I did a lot of investigations up to now and figure out that the cap of my sensor or sensor itself were not correctly soldered, glued or some mechanical stress was applied. After some cleanup of the IC head and window and re glueing the black plastic cap I got more stable operations but again when moving object stable readouts between ca 36 - 80 mm, when stop object stable readouts from 80 - 250mm for short mode. I took the second dk-ch101 board and this one is much stable the previous at least no problems with sensor caps or glueing processes. The 2nd board is showing the same results 36-80 mm stable only when moving objects, 80-250 mm stable when object is not moving. I have used both fabrik firmware and SonicLink software as well as HelloChirp example and Atmel Studio 7 enviroment. Could You provide Your email address or skype nick (my skype is konrad6155, my email highvoltagecontrol@gmail.com). I would like to exchange experience with You (if possible ofcourse0 and again get the same working point as You have, I mean 2-40mm range stable operation for none moving objects. It looks like forum is very limited with possiblities to exchange files and udpate rate between our posts is quite long (1-2 days). I wrote to support about this issue but looks like nobody is answering so far.

phpbb Post ID

38353

I did more investigations today and it looks like amplitude detection for very short distance below 80 mm is not stable (look to my pdf first two amplitude plots obtained from IQ for the same distance of 36 mm), the peak of amplitude is the same sample position but 3 times less level when object stop then when it was moving, for the stationary objects above 80 mm the amplitude peak is anyway relatively small but from measurement to measurement quite stable, i did shots for 100 mm, 120 mm, 140 mm, could be few reasons for the problem: 1) firmware bug inside ch101 dsp processor, 2) firwmare bug inside sam55 processor or this is the real limitation of the ch101 IC, it is just not stable for short distances and stationary objects. I do not expect hardware problem since for longer distances above 80 mm and normal distance up to 1000 mm all is fine and quite stable.

Concerning schametics of dk-ch101 and ultrasonic boards there are 2x i2c buses are first translated usign 3.3V to 1.8V translator, then i2c#0 is muxed to onboard populated sensor0 or external corner FFC sensors 0,1, i2c#1 is directly connected to corner FFC sensors 2,3. I would not expect i2c muxes switching errors since I have only middle one (populated) and it is correctly discovered, others not (i have ordered FFC tapes for them).

Echo signal amplitude must come from other fact I would expect firmware bugs or IC limitations itself. Lets exchange more info on skype or via emails (just send me Your details)

phpbb Post ID

38354

We use 100V peak voltage bursts and a 60dB amp to sense 1m with a 200kHz ultrasonic sensor. The ch101 has 3.3V for bursting, so I think they have massive dsp which is able to sense any motion inside the noise and static measurements are simply not possible. I see the same effect: sensing motion and after stopping motion the echo is vanishing.
And yes, this forum site is annoying.
Regards

phpbb Post ID

38510

Guys it's been a while you created this topic but I wanted to give my opinion for anyone who will see this topic later. If you enable the STR(Static Target Rejection), it will only detect the moving objects. It is mostly used for presence detection in the rooms etc. In your case, if you say that your device can only detect the moving objects, the reason could be that you enabled STR. You should disable it if you want to detect everything.

Regards.

phpbb Post ID

38791

Thanks for the advice. I am using
(Console output from Smartsonic)
Using class 'SmartSonic.Sensor_CHRP101' from C:\Eig_SW\SmartSonic Platform 1.11.1\data\CH-101\Sensor_CHRP101.cs
Doing Init
Init CH-101
Sensor 0 operating frequency is 177418
Sensor 0 IQ sample size is 39
A scan selected
Init Done
Logging 'CH101 0' started
Logging 'CH101 0' paused
Logging 'CH101 0' started
Sensor #0 closed

with

            public BoardSettings()

                : base("CHIRP 101")

            {

                FSR_Options = new int[] { FS_3, FS_4, FS_5 };

                FSRSR_Options = new int[] { FS_0, FS_2, FS_6};

                SamplingRate_Options = new int[] { SR_0, SR_1, SR_2, SR_3, SR_4, SR_5 };

                ASCAN_SamplingRate_Options = new int[] { SR_0, SR_1, SR_2, SR_3 };

                SensorMode_Options = new SensorMode[] { SensorMode.Normal, SensorMode.ShortRange , SensorMode.Narrow_FOV, SensorMode.Narrow_FOV_SR };

                SensorOperating_Options = new SensorOperationType[] { SensorOperationType.Not_Selected, SensorOperationType.Tx_Rx, SensorOperationType.Rx_only };

                FSRSelected = FSR_Options[0];

                FSRSRSelected = FSRSR_Options[0];

                SamplingRateSelected = SamplingRate_Options[0];

                ModeOfOperation = SensorMode_Options[0];

                OperatingType = SensorOperating_Options[0];

                SensorModel = 1;

		        ThresholdSupported = false;

                STRSupported = true;        //this makes the checkbox visible

                IQSize = 150;

            }

inside the cs file and I cannot get a difference between checked/unchecked STR checkbox.
The target @ 50mm distance generates a 5000digits echo, when it is moved and after a while when rested the signal goes below 1000#.

What else should I do to get everything?

Regards.

phpbb Post ID

38800

phpbb Topic ID

38308