Hi all. I recently purchased an K8031 from a local supplier. Everything went fine in terms of assembly and construction. The calibration also went OK. Now the problem is that there are some spikes on the trace. They are independent of the fact that the probe is shorted or that I apply any input signal. When doing the calibration I am requested by the instruction manual to connect the probe to the J4 test point. The trace is OK, I am able to adjust the middle position of the trace, also the 2.50V amplitude, but the spikes appear on the up and down levels of the measured signal (not on the rising or falling edge). Generating and studying the text files I noticed that there is a slight oscillation in the value of the measured signal even if that signal has no amplitude modification. The same thing happens even if there is no signal applied on the probe (probe shorted), there are 1 bit high variations for example (sample from the text file generated):
The problem gets even more complicated when selecting a faster timebase, for example 0.1us (notice that now the spikes are much higher, I measure 6V when in fact there is no input signal):
At first I thought that there may be a communication problem. I changed the computer. No go. I tried powering the output optocoupler stage from a different more stable power supply (thinking that the LPT port might not be able to deliver enough current). Again no result. Tried removing the input opamp from the socket (thinking that there may be some small signal interfereces, or that the trace positioning mechanism may be introducing some noise in the input stage). Again nothing. Tried removing the sockets from the opamp and the A/D converter and soldering them directly on to the board. No way.
After all the thinking and working I realized that it may be an A/D converter issue. Now, I know that TDA8703 is not an excellent A/D converter, but I still hope that it may be an implementation problem and there is a solution for this.
I hope I gave you enough information for you to understand what is happening. If not, feel free to ask, I will provide anything that is needed for solving the problem.[/url]
The random 1 bit variation is normal. It is caused by the operation of the AD converter and input amplifier noise.
The images you got at higher sample rate are not OK.
Is there all the time the spikes as in the picture you sent?
Does the spike appearance change if you change the trace position in vertical direction on the screen?
Somehow the first answer to my problem had been lost. Someone gave me suggestions about placing saome capacitors to the ADC pins. I would like to ask that person to answer me again with those suggestions as I did not save them. Thank you :shock:
Yes the spikes change if I change position of the trace on the screen.
As you assumed this may be some sort of ADC problem.
You may try to improve the supply voltage filtering of the TDA8703.
Add 10 nF ceramic capacitor with very short wires (1 mm) directly between the pins 19 and 20 of the TDA8703. If no help solder another 10 nF between pins 21 and 20.
If no help you have to change the TDA8703.
Please read the instructions how to proceed: velleman.be/be/en/support/warranty/
There is no improvement by adding the two 10 nF capacitors that you suggested. Any other ideas? :roll: :?:
The warranty solution is somehow complicated because I have to send the whole kit to my supplier, wich is not in my town, convince them that there is a problem with it, then they send it to velleman, then velleman gives a solution, then send it back to my supplier, wich in turn send it back to me. It is indeed a complicated and long time consuming procedure. Is there any way that I can receive a replacement for TDA7803 directly from velleman? Do I have to send the defective part to them?
I still hope that there is a simpler solution. I don’t think that I am the first to have this kind of problem. Anyone, any ideas?
Could you please post a link to the image showing the problem of your PCSU1000 scope. The current link is to the K8031 image posted by edugeo.
You may use some free image hosting services (e.g. servimg.com/) to upload the .PNG image and to get a link to it. In PCSU1000 select “File/Save Image” to save the image in lossless, compressed PNG format.
It seems that you measured the test signal from the rear connector of the scope. As you see, that signal is not very “clean”. It is intended only for the x10 probe compensation setting.
When you measure “normal” signals I think there is no ripple like seen in the image you sent.
If possible use x10 setting of the probe.
Here an example of signal from PCGU1000 function generator:
BTW: Use the average option to reduce the noise. In this image it is not used.
Indeed, the probe was 1/10 measuring the calibration point on the rear of the bezel.
As I can see, you have almost the same result on your screenshot.
So it seems that the ripple is enclosed into this scope interface.
I must say, I’m not very happy with these result. My old 20 MHz scope performed better.
