Baracuda Build (silverdog kit Rev2)

[Koala]

Good progress. Still don't see why some work and some don't. I take it my Earth field is working correctly with no modifications. Those advertising fidget magnets made out of magnetic paper have no affect on the threshold. Is this a valid test ??

[Qiaozhi]

Hi Q,

That is good news☺. Do you have a new value of the 150K resistor?

Ray

The 150k resistor sets the Earth field pulse delay, so I just left it as is. You can reduce the delay (if you like) by lowering this value. However, there will probably not be any detectable difference in operation.

[raygdunn]

Hi Q,

That is good news☺. Do you have a new value of the 150K resistor?

Ray

[Old cart]

At last I think I understand the problems with the Baracuda ...
Impulse's Baracuda PCB is now working as I would expect, including Earth field elimination.

Here's what I believe has happened over time:
Initially the design was intended to run on a nominal 9V battery pack. i.e. 6x 1.5V AA batteries (or 8x 1.2V NiCd batteries). The schematic indicates a voltage range at the input of 8.4V to 12V. These limits are imposed by the 7660S (or 7660A), as 12V is the maximum voltage input allowed, and [if I remember correctly] 8.4V is about where the 7660 tends to drop out.

In this case, if you run the Baracuda with a 12V battery it will most likely have a higher terminal voltage, and could quite likely destroy the 7660. Actually, if you read the datasheet carefully, it does state further on that the 7660 has an absolute maximum rating of 13V, so you might get away with it for a limited time. One possibility here would be to fit a [pin compatible] LT1054, which has a maximum input voltage of 15V.

I also suspect that the original design was using + and -5V on the CD40106 (although it is difficult to read in the original schematic) and someone noticed that this caused the gate voltage of the sampling FETs to go positive. This was when the supply was changed to 0V and -5V, but whoever made the change forgot to also change the pull-up resistors from +5V to 0V. Instead they changed the pull-up resistor values to get back the original pulse widths. The problem here is that (even if you manage to avoid blowing up the 7660 with too high an input voltage) the CD40106 can go kaput due to the capacitors in the power supplies being incorrectly connected to the positive terminal of the battery.

The screw-up with the power supply capacitors appears to have been there from the very start. Consequently I have removed the 220uF capacitor at the input of the 78L05, and fitted this at the output (referenced to 0V). The 220uF capacitor at the 79L05 output has also been changed to be referenced to 0V. There should have been a capacitor at pin5 of the 7660, which feeds the 79L05 (you will see it's missing on the schematic), and there was no capacitor present. I added a 10uF cap to 0V. This boosted the 7660's negative output voltage from about -6.8V to be about -9V (my supply was set at 10v). A 1V drop on a 7660 is quite normal for this device. With only -6.8V at the input of the 79L05, this was affecting the negative voltage supply to the CD401016. Also, I removed the 220uF capacitor between pins 2 and 4 of the 7660, and replaced it with a 10uF.

By the way, it does indeed appear that the 2M2, 470nF and 1N4148 are synchronising the 7660 to the TX oscillator. This is similar to what the MPP does, in that it allows the voltage converter to use its [faster] internal oscillator, but causes it to pause momentarily during the TX pulse. Hence the internal oscillator gets locked to the TX. You can see this yourself by probing the TX oscillator with the scope using CH1, and then probing pin2 of the 7660 with CH2. The internal oscillator frequency should be around 10kHz, but it appears to closer to 7.8kHz. Apparently you can adjust the internal frequency by changing the value of the 10uF capacitor between pins 2 and 4, so it might be worth checking that out. However, I would be surprised if it would need increasing to 220uF as before.

Because moving the pull-up resistors on the CD40106 to 0V alters the pulse widths, I had to change the 33k on pin 1 to 15k. This reduced the 100us main sample to a more sensible 46us. I then removed the 33k resistor on pin 3, and temporarily connected a 50k pot. As you would expect, the Earth field was eliminated almost entirely when the main and secondary sample were pulse widths were equal. This occurred when the pull-up resistor on pin 3 was 12k. The pot was then replaced by a 12k fixed resistor.

When I can get some time, I'll create a new schematic incorporating all the changes.

Splendid work Q. I knew you would be successful! I suspected many of these flaws but had no board to work on but YOU DID IT!

Thanks!

[Qiaozhi]

At last I think I understand the problems with the Baracuda ...
Impulse's Baracuda PCB is now working as I would expect, including Earth field elimination.

Here's what I believe has happened over time:
Initially the design was intended to run on a nominal 9V battery pack. i.e. 6x 1.5V AA batteries (or 8x 1.2V NiCd batteries). The schematic indicates a voltage range at the input of 8.4V to 12V. These limits are imposed by the 7660S (or 7660A), as 12V is the maximum voltage input allowed, and [if I remember correctly] 8.4V is about where the 7660 tends to drop out.

In this case, if you run the Baracuda with a 12V battery it will most likely have a higher terminal voltage, and could quite likely destroy the 7660. Actually, if you read the datasheet carefully, it does state further on that the 7660 has an absolute maximum rating of 13V, so you might get away with it for a limited time. One possibility here would be to fit a [pin compatible] LT1054, which has a maximum input voltage of 15V.

I also suspect that the original design was using + and -5V on the CD40106 (although it is difficult to read in the original schematic) and someone noticed that this caused the gate voltage of the sampling FETs to go positive. This was when the supply was changed to 0V and -5V, but whoever made the change forgot to also change the pull-up resistors from +5V to 0V. Instead they changed the pull-up resistor values to get back the original pulse widths. The problem here is that (even if you manage to avoid blowing up the 7660 with too high an input voltage) the CD40106 can go kaput due to the capacitors in the power supplies being incorrectly connected to the positive terminal of the battery.

The screw-up with the power supply capacitors appears to have been there from the very start. Consequently I have removed the 220uF capacitor at the input of the 78L05, and fitted this at the output (referenced to 0V). The 220uF capacitor at the 79L05 output has also been changed to be referenced to 0V. There should have been a capacitor at pin5 of the 7660, which feeds the 79L05 (you will see it's missing on the schematic), and there was no capacitor present. I added a 10uF cap to 0V. This boosted the 7660's negative output voltage from about -6.8V to be about -9V (my supply was set at 10v). A 1V drop on a 7660 is quite normal for this device. With only -6.8V at the input of the 79L05, this was affecting the negative voltage supply to the CD401016. Also, I removed the 220uF capacitor between pins 2 and 4 of the 7660, and replaced it with a 10uF.

By the way, it does indeed appear that the 2M2, 470nF and 1N4148 are synchronising the 7660 to the TX oscillator. This is similar to what the MPP does, in that it allows the voltage converter to use its [faster] internal oscillator, but causes it to pause momentarily during the TX pulse. Hence the internal oscillator gets locked to the TX. You can see this yourself by probing the TX oscillator with the scope using CH1, and then probing pin2 of the 7660 with CH2. The internal oscillator frequency should be around 10kHz, but it appears to closer to 7.8kHz. Apparently you can adjust the internal frequency by changing the value of the 10uF capacitor between pins 2 and 4, so it might be worth checking that out. However, I would be surprised if it would need increasing to 220uF as before.

Because moving the pull-up resistors on the CD40106 to 0V alters the pulse widths, I had to change the 33k on pin 1 to 15k. This reduced the 100us main sample to a more sensible 46us. I then removed the 33k resistor on pin 3, and temporarily connected a 50k pot. As you would expect, the Earth field was eliminated almost entirely when the main and secondary sample were pulse widths were equal. This occurred when the pull-up resistor on pin 3 was 12k. The pot was then replaced by a 12k fixed resistor.

When I can get some time, I'll create a new schematic incorporating all the changes.

[Old cart]

It's not actually an xray it's a hand drawn vector graphics image so that we can enlarge without blur... I could blame my eyesight or old age but I simply missed it...
Here's an update...
[ATTACH]35824[/ATTACH]
...

Slightly off topic...

In the process of building a little tool to allow me to generate pulses for TX and Delays using ATMEGA328...
It started life as a front-end to a simple metal detector but after changing several resistors to get the right timing on my bara, I thought I'd use it as a tool before I screw up the bara PCB...

The aim was simple, I wanted to generate an exact 100uS pulse for the coil at about 600 PPS, followed by a delayed sample pulse of 45uS and finally the second sample pulse.
By using a microcontroller, a 16/20character LCD and four push buttons, I can vary all the pulses and delays in 5uS steps.
[ATTACH]35829[/ATTACH]

The basic design is complete but I haven't started on the code yet... still it might be useful...
Mike

That new drawing is fine. Thank you so much
That pulse generator looks interesting keep us appraised about progress.

[Michaelo]

It's not actually an xray it's a hand drawn vector graphics image so that we can enlarge without blur... I could blame my eyesight or old age but I simply missed it...
Here's an update...

...

Slightly off topic...

In the process of building a little tool to allow me to generate pulses for TX and Delays using ATMEGA328...
It started life as a front-end to a simple metal detector but after changing several resistors to get the right timing on my bara, I thought I'd use it as a tool before I screw up the bara PCB...

The aim was simple, I wanted to generate an exact 100uS pulse for the coil at about 600 PPS, followed by a delayed sample pulse of 45uS and finally the second sample pulse.
By using a microcontroller, a 16/20character LCD and four push buttons, I can vary all the pulses and delays in 5uS steps.


The basic design is complete but I haven't started on the code yet... still it might be useful...
Mike

[raygdunn]

Hi Old cart
5534 Pin 4 is OK. It is connected to -5V via. two top of board links going right then down in the board top view. Plus a 100nF to earth. Or have I mis-understood?

Ray

Ah I missed that.

Sorry I'm wrong.

Ray

[Polymer]

Note! this X-ray and Polymers lack a connection to pin4 on the 5534. I don't know what is going on here. It appeared in earlier X-ray posts like 57.

@Old Cart: You are right!
@Michaelo: I assume you did a complete re-draw for the vector layout and missed that little track. No worries. Could you please put it in and post. Thanks.

I'll be travelling the next few weeks, also detecting , but will only have simple computing possibilities on the way, like writing ...

Here's a closeup of both:

[raygdunn]

Note! this X-ray and Polymers lack a connection to pin4 on the 5534. I don't know what is going on here. It appeared in earlier X-ray posts like 57.

Hi Old cart
5534 Pin 4 is OK. It is connected to -5V via. two top of board links going right then down in the board top view. Plus a 100nF to earth. Or have I mis-understood?

Ray