Voltage on my coil is 40V, not 350V

[dotniet]

Thanks CAS.
I put together last night 10 rechargeable batteries. They are about 2Ah each. I hope this will do it. I will try it tonight.

Since the RX needs large current only short time, a large capacitor on my power supply should also work. But I will try the battery pack first.

[dotniet]

Ok. I replaced the transformer with a battery pack. It has about 12.5V, 2Ah. I also added a 2200uF capacitor but it makes no diff.
But nothing changed after switching to battery pack. The voltage on coil is the same (160v).

Why can't I get 350V on coil and 58uS pulse width???


-----------------



TP1:
Frequency: 0.93ms (instead of 1mS)
Pulse width: 49uS (instead of 58uS)




TP3:
1.5v drops below 1v for a 15cm steel pot top


Voltage on coil is now 160V.



Step 6:
I still cannot adjust the distance between TX and main sample. This is the wave form with the trimmer at min and at max:

[Qiaozhi]

The flyback voltage level is most likely not being displayed correctly by your PC scope due to lack of bandwidth. Clearly there has already been an improvement from 40V to 160V by switching over to a battery pack, so I would ignore that as a potential problem. Since the TX pulse width is also lower than expected, the accuracy of your scope measurements appear to be suspect.

As far as the main sample delay is concerned, this is working correctly. TP5 should show a pulse that changes in width as you adjust R6, The main sample pulse can be monitored on TP7. When you adjust R6, the pulse position (relative to the TX pulse) will change. Have a look at Fig. 10 in the Build Document.

[dotniet]

The flyback voltage level is most likely not being displayed correctly by your PC scope due to lack of bandwidth.

My oscilloscope is SainSmart DDS140. Bandwidth 40MHz. 78euros.

Clearly there has already been an improvement from 40V to 160V by switching over to a battery pack

Small correction. The voltage on coil was 40V because I had the wrong coil. After installing the correct coil the voltage increased to 160V.
After switching from transformer to battery pack nothing changed.

Since the TX pulse width is also lower than expected, the accuracy of your scope measurements appear to be suspect.

You may be right. I don't know about the hardware but the manual is simply sh*tty and the software is not much superior. They didn't even provided the correct manual on the CD.

As far as the main sample delay is concerned, this is working correctly. TP5 should show a pulse that changes in width as you adjust R6, The main sample pulse can be monitored on TP7. When you adjust R6, the pulse position (relative to the TX pulse) will change. Have a look at Fig. 10 in the Build Document.

I GOT IT NOW!!! I followed the manual (build document) EXACTLY and this caused the problem. The manual says that I should monitor TP1 and TP5 and (without mentioning TP7) I should adjust R6 until I get the right delay. But the delay is measured on TP7 (which again is not mentioned in that paragraph.

Monitor the TX oscillator at TP1 using an oscilloscope (channel 1), and TP5 (channel 2). Trigger off channel 1. Short pin2 of PL3 to either pin 1 or pin 3, and adjust R6 for a minimum sample delay of 22us.

You may want to adjust or rephrase this in the manual for other novices like me

Thanks for clarifying this. I think I am on the correct track again! Thank again!!!!!!!!!!!!!!!!!!!!!!

I hardly wait to go back and install the rest of the parts!!!!
And thanks one more time.

[Qiaozhi]

I GOT IT NOW!!! I followed the manual (build document) EXACTLY and this caused the problem. The manual says that I should monitor TP1 and TP5 and (without mentioning TP7) I should adjust R6 until I get the right delay. But the delay is measured on TP7 (which again is not mentioned in that paragraph.

The main sample delay is measured at TP5, whereas the main sample pulse width is measured at TP7.
Perhaps you're not used to the PI terminology, but the Build Document is correct.

I wouldn't be too concerned about the flyback voltage, as your mileage may vary depending the coil construction.

Glad to hear that you're now back on track.

[dotniet]

The main sample delay is measured at TP5

Ok, now I am confused.
So, if I hook the oscilloscope at TP1 and TP5 and adjust the R6, I should be able to change the delay between the two signals I see on screen. Right???
But when I adjust R6 I see no change in the delay between the signals!
(I am speaking strictly about TP1-TP5 here.)

This is what I was showing in my previous screenshots: as I adjust R6, the yellow signal (at TP5) changes its WIDTH but NOT its DELAY (distance to blue/TP1) signal:





What I don't understand?
Where am I wrong?

Well, it might be that I am using the oscilloscope wrong (I used an oscilloscope 15 years ago. I purchased this one 1 week ago especially for this project).
The trigger is set to channel 1 (blue).

But anyway, if you say that what we see in the screenshots is right and I can proceed to Step8, I will do so. You are the guru here.

[CAS]

Just remember that TP5 is just the clock input for U6b and U7a. TP5 does not show the delay in reference to the TX pulse /main clock pulse. The width of the pulse at TP5 is the delay. Varying R6 will alter the width of the pulse at TP5 therein varying the delay. Have a look at the circuit and see where TP5 is located and you will see what I mean.

[6666]

dotniet



You may be having the same problem with your MPP as mine.
When I first set mine up according to the build doco , by triggering on TP1 and measureing the delay on TP5 it was easy to adjust the delay with R6, in fact I published a picture
but it appears some fault is now on my PCB and no amount of adjusting R6 affects the delay, there is no delay at all, very strange, this needs more investigation, but not right now for me.

[Qiaozhi]

You may be having the same problem with your MPP as mine.
When I first set mine up according to the build doco , by triggering on TP1 and measureing the delay on TP5 it was easy to adjust the delay with R6, in fact I published a picture
but it appears some fault is now on my PCB and no amount of adjusting R6 affects the delay, there is no delay at all, very strange, this needs more investigation, but not right now for me.

dotniet does not have a problem with his MPP. It is working correctly. Your issue is being caused by a either a component failure, an unintended short, or perhaps a wiring error.

CAS has already explained how the timing generators work, but I will reiterate here:
The MC14538BCP ICs are monostable multivibrators, which are used to generate a single pulse of a defined width when a negative-going pulse is applied to the CLK~ (B) input. U6a is triggered by the TX oscillator so that, when the MOSFET is turned off, a pulse is generated at the Q output of U6a with a width that can be set by adjusting R6 (minimum sample delay) and the REJECT pot (when fitted) to set the sample delay when in use. This is the main sample delay (TP5). This signal is then fed into the CLK~ (B) input of U7a, which triggers the monostable on its negative-going edge to produce another [delayed] pulse at TP7. This is the actual main sample pulse, which has been delayed by the preceding pulse at TP5.
The main sample delay pulse (TP5) is also fed into U6b, which generates the secondary sample delay (set by R10 and C10), and can be monitored at TP6. This pulse triggers U7b to generate the secondary delay pulse. The components C11 and R11 set both the main and secondary pulse widths to the same value, with the help of D7 and D4. This is because the main and secondary pulse widths must be the same width for effective Earth Field cancellation to occur.

This is all shown very clearly in Fig. 10 on page 15.

[dotniet]

Just remember that TP5 is just the clock input for U6b and U7a. TP5 does not show the delay in reference to the TX pulse /main clock pulse. The width of the pulse at TP5 is the delay. Varying R6 will alter the width of the pulse at TP5 therein varying the delay. Have a look at the circuit and see where TP5 is located and you will see what I mean.

Hi CAS.
That was what I failed to understand (that TP5 IS the delay width itself).
In the end I figured out a different way to measure all the signals: I measured one signal relative to the previous one. More exactly main sample to TX signal and second sample to main sample.

Thanks for clarifying what TP5 actually does.

---------------------------------------------------------

So, except the fact that I calculated the coil wrongfully, everything went smooth.
Now the detector is ready.
I tested it for about 10 minutes. Yes, the joy is BIG but it is late and I am tired. So, I will test it more tomorrow.
For now I can say this: it detects a 1EURO coin at a distance of 21-25cm and the cola (beer, couldn't find cola) can at about 50 cm.

Thanks you all for support!

[sinclairuser]

i just wanted to add, your flyback voltage will be higher than 160, the way the sainsmart 140 works the 10x probes are not mega accurate(for ac flyback voltages), with 100x probes and proper calibration however they can be excellent.
as george said dont worry about it, it appears to work fine, and well done on the choice of scope, very good and value for money.

ac voltages is where these usb pc scopes sometimes fall down, but this particular one is well rated 50v 1x, 500v 10x the 100x probes just give it the edge.
i used to recommend an older analogue scope for beginners, but with these its not worth the gamble, buying second hand can be a minefield now at 76 quid why bother, just get a sainsmart 140.

[dotniet]

but this particular one is well rated 50v 1x, 500v 10x the 100x probes just give it the edge.

DDS 140 is rated 5V for 1x probe and 50V for 10x probe.
No 100x probe provided.

[sinclairuser]

DDS 140 is rated 5V for 1x probe and 50V for 10x probe.
No 100x probe provided.

i got the probes mixed up 500v is for the 100x probe, you dont it get with it, you have to buy them separately from sainsmart.

[dotniet]

and well done on the choice of scope, very good and value for money.

Yeah... The price is really small... but what is this good for when the oscilloscope measures 160V instead of 350V ?
That's a pretty damn high error.

[Qiaozhi]

Yeah... The price is really small... but what is this good for when the oscilloscope measures 160V instead of 350V ?
That's a pretty damn high error.

When the MOSFET is switched off, the dV/dt of the flyback pulse is huge. So it could quite simply be that the scope does not have enough resolution for such a fast signal. The sample rate is quoted as 200M Sa/s, which equates to a sample every 5ns, so it may just be missing the peak of the flyback. My own scope has a bandwidth of 25MHz, but a sample rate of 250M Sa/s, and [by decreasing the timebase] you can see that the flyback voltage varies from pulse to pulse. This is, of course, an anti-aliasing effect from the sampling process.

Personally, I wouldn't worry about it. What's happening with the coil current is more important than the flyback voltage, which is an unwanted side-effect. The only thing about flyback voltage, that you need to be concerned about, is whether it causes the MOSFET to go into avalanche mode. This will cause the MOSFET to get hot, and can lead to premature component failure.