Baracuda Build (silverdog kit Rev2)

[Polymer]

@ Qiaozhi

Hi Qiaozhi, I think i remember reading that you are noting/collecting infos on a revision 3 for the baracuda.

My Revision 3 baracuda board wish/suggestion list.
Some things may already have been mentioned by others.

PCB Related:
- Bigger holes in general (like Surf PI)
- Big holes for: power supply wires, coil wires, charge capacitor for coil, Rdamp
- Some holes for charge capacitor in axial direction and/or more holes to place two axial/radial ones in parallel.

If possible space wise:
- Multiple pads/holes for different component sizes (especially box caps, caps)
Will cost more for pcb, but getting box caps in right size may be difficult for some, bending the short
pins to suit doesn't work well.

- How about power supply wires entering PCB close to charge capacitor?


Regarding the 7660:
It looks like roughly 1/2 to 2/3's on offer on the market are 10 volt versions.

In the datasheets 10uF capacitors are usually called for with
100uF ones being mentioned to get lower esr for better effectivity.
Are the 220uF capacitors really necessary?

PIN 7, Oscillator. The 2M2 resistor still has me puzzled
and I question if this setup on the baracuda really works, or is it a special tweak?

I read into several datasheets and came to the conclusion: Different manufacturers, different frequencies.
Since Pin 7 can also work to change frequency with changing of capacitance from 1pF thru to 1000pF,
MIGHT the 2M2 setup somehow be mimicking a variable capacitor?

Looked at the PSU in Surf PI. It has one voltage regulator less.
Advantages/disadvantages? Could it be used for the baracuda?

Thanks for your patience!

It looks like something is being done with the 40106, which is great. I am looking forward to this and maybe other
improvements that the electronically adept are working on.

[Polymer]

I started off making the rev 2 but ended up adding the recent fixes and kept on going... but your right, it will confuse so I will edit and post rev 2 board...
I may add one with all official fixes later...

I understand how easily things like that can happen. I might have been a bit strict, but I started with the Baracuda 1 1/2 years ago and found reading all the posts quite confusing and put it to the side.
I would like newcomers to not have to go through the same experience, although I did learn a lot on the way which I did benefit from.
As I see it you have kicked off the development of revision 3 with your thread and I am pleased to see that.
Looking forward to having your improvements added to the new version.

[Michaelo]

Nice vectored rebuild Michaelo!

What's the light green track doing in there?

Please don't put anything new on the board ... yet. It may cause more confusion.

As 6666 has mentioned could you please correct the typo "LF421" to "LF412".

Thanks!

I started off making the rev 2 but ended up adding the recent fixes and kept on going... but your right, it will confuse so I will edit and post rev 2 board...
I may add one with all official fixes later...

[Old cart]

KT315, while nicads can reach high voltages right after charge, they lose about 1% of their charge per day just sitting there. I find they stabilize at about 1.25 volts per cell rather quickly. When they reach 1.1 volts per cell they are considered to be discharged.
That having been said it is wise to design the circuit to withstand at least 1.5 volts per cell. I also like to allow a 20% safety margin.
Here is one of the most useful documents I have seen regarding charging nicads. Note this site also addresses many other battery chemistries.

[Old cart]

Here you go ...

[ATTACH]35755[/ATTACH]

Even better, this is great. Thank you.

[kt315]

KT, I have a few NiCd cells here at home along with a whole lot more NiMH, if that 1) somehow changes the way how these cells behave, or 2) makes you read anything I write without spite-tinted glasses, or 3) stops you changing arguments and voltage fractions by editing your previous posts.

The reason why I brought up chargers and recovery after charging was personal experience in characterizing a number of cells for products at the workplace as part of sourcing phase. Different chargers, testbox temperatures and loads for each cell under test, with a few supermarket brand cells in for comparison. I don't have the graphs any longer as it was some years ago, they'd probably fall under joint NDA, but here's something from a decent manufacturer. http://www.farnell.com/datasheets/86307.pdf

They show the same kind of cell voltage vs discharge time curve, with an initial full cell peak and a steady descent over a noticeably long time to a gentle plateau, somewhere tenth of a volt around the god's gospel 1.2V depending on discharge rate. At the higher discharge rates test, the initial cell voltage is recorded at 1.45V which could be due to shorter averaging period in their test gear, or a very full cell to get a nice peak out. Another noteworthy thing in the datasheet is the less often shared cell voltage vs. charge time curve, where we see that high peak voltage in fast charging.

Either way, that peak voltage in a fully charged cell still stands out there, and overvoltage damage in semiconductors is generally lightning fast and final. Those 10V rated charge pumps are more than likely the same manufacture as 12V rated ones that just didn't consistently survive 12 volts at the bench but could be safely rated at 10. When teaching electronics it is an often shared rule of thumb to try and overspec your parts 50% or more, to extend their lifetime; we generally see 35V electrolytic caps instead of 25V ones on a 24V rail even if that rail was regulated within 100mV.

Apologies to Koala and the rest, I believe we should move this to General Electronics under nickel battery behavior ... if we really need to keep dragging this on, KT.

ok. try charge your nicad and show me the voltage by camera pics. i want to see '1,5 volts'.

[Polymer]

Though I play around with Xara (vector graphics) and put together an image inspired by Polymer's xray image...
[ATTACH]35748[/ATTACH]

Don't appear to be able to attach as inline... ;(

I like to play with vector graphics and I thought why not... also as my eyesight is poor, being able to enlarge without loss of composition is a bonus... mind you I still haven check every single component yet... I've added the 40106 resistor fix... it might come in useful if you need really large images without blurring...
Mike

Nice vectored rebuild Michaelo!

What's the light green track doing in there?

Please don't put anything new on the board ... yet. It may cause more confusion.

As 6666 has mentioned could you please correct the typo "LF421" to "LF412".

Thanks!

[Polymer]

Very nice! I to have bad vision. Is there any way to change the links on the top of the board to a different color? Red blends in. Better yet make the board gray or black.

Here you go ...

[Qiaozhi]

In the final version it may be worth making the EFE adjustable.

[Old cart]

EF elimination adjust

Hi raygdunn

Thanks for your input
I have looked into this and tried different coils and different cables to the coil and also changed the case, no of this made any difference

While using the Barra at the beach,
At the end of the swing, I tend to twist and lift the coil a little
If I pay more attention to the end of the swing by not twisting or lifting, the barra is fine

I am thinking that it is more to do with Earth field elimination like Qiaozhi suggested


Hi Qiaozhi

Is there a way that I can test and fix the earth field elimination without an oscilloscope?

Any and all help is greatly appreciated

An earth field elimination problem is entirely possible. The main and earth field pulses need to be the same width. The problem is the 40106 is anything but a precision device. With a scope it is easy to adjust the two sample pulse widths by varying resistance on the main and EF sample. Without a scope it is tougher. I suppose it is possible to wave a small ceramic or refrigerator magnet near the coil while adjusting the EF width. Adjust for minimal response to the magnet. Do not use an iron magnet, the detector will detect it no matter here the Ef cancel is.

If you want to try this put a 100k pot in series with a 33k fixed resistor. Adjust the combo to max resistance, should be about 133k. Then solder the whole mess across the 33K resistor connected to pin 3 of the 40106. As you slowly adjust the pot while waving the magnet you should, if you are lucky, find a point where there is minimal response. If you do not it probably means this pulse was too narrow to begin with or something else is wrong.
The quickest way to isolate this is to remove the existing 33k resistor and replacing it with the pot/33k combo. Repeat the procedure above. If it does not work this time, something else is wrong. You will probably need a scope.
if it does work then remove the combo careful, without disturbing the setting, and measure the combined resistance. Select a fixed resistor very close to the value and add it to the board in place of the original resistor.

[6666]

Though I play around with Xara (vector graphics) and put together an image inspired by Polymer's xray image...
[ATTACH]35748[/ATTACH]

Don't appear to be able to attach as inline... ;(

I like to play with vector graphics and I thought why not... also as my eyesight is poor, being able to enlarge without loss of composition is a bonus... mind you I still haven check every single component yet... I've added the 40106 resistor fix... it might come in useful if you need really large images without blurring...
Mike

Whats an LF421

[Old cart]

Though I play around with Xara (vector graphics) and put together an image inspired by Polymer's xray image...
[ATTACH]35748[/ATTACH]

Don't appear to be able to attach as inline... ;(

I like to play with vector graphics and I thought why not... also as my eyesight is poor, being able to enlarge without loss of composition is a bonus... mind you I still haven check every single component yet... I've added the 40106 resistor fix... it might come in useful if you need really large images without blurring...
Mike

Very nice! I to have bad vision. Is there any way to change the links on the top of the board to a different color? Red blends in. Better yet make the board gray or black.

[Paul B]

Hi raygdunn

Thanks for your input
I have looked into this and tried different coils and different cables to the coil and also changed the case, no of this made any difference

While using the Barra at the beach,
At the end of the swing, I tend to twist and lift the coil a little
If I pay more attention to the end of the swing by not twisting or lifting, the barra is fine

I am thinking that it is more to do with Earth field elimination like Qiaozhi suggested


Hi Qiaozhi

Is there a way that I can test and fix the earth field elimination without an oscilloscope?

Any and all help is greatly appreciated

[Michaelo]

Rev 2 (vectored)

Though I play around with Xara (vector graphics) and put together an image inspired by Polymer's xray image...


Don't appear to be able to attach as inline... ;(

I like to play with vector graphics and I thought why not... also as my eyesight is poor, being able to enlarge without loss of composition is a bonus... mind you I still haven check every single component yet... I've added the 40106 resistor fix... it might come in useful if you need really large images without blurring...
Mike

[ODM]

KT, I have a few NiCd cells here at home along with a whole lot more NiMH, if that 1) somehow changes the way how these cells behave, or 2) makes you read anything I write without spite-tinted glasses, or 3) stops you changing arguments and voltage fractions by editing your previous posts.

The reason why I brought up chargers and recovery after charging was personal experience in characterizing a number of cells for products at the workplace as part of sourcing phase. Different chargers, testbox temperatures and loads for each cell under test, with a few supermarket brand cells in for comparison. I don't have the graphs any longer as it was some years ago, they'd probably fall under joint NDA, but here's something from a decent manufacturer. http://www.farnell.com/datasheets/86307.pdf

They show the same kind of cell voltage vs discharge time curve, with an initial full cell peak and a steady descent over a noticeably long time to a gentle plateau, somewhere tenth of a volt around the god's gospel 1.2V depending on discharge rate. At the higher discharge rates test, the initial cell voltage is recorded at 1.45V which could be due to shorter averaging period in their test gear, or a very full cell to get a nice peak out. Another noteworthy thing in the datasheet is the less often shared cell voltage vs. charge time curve, where we see that high peak voltage in fast charging.

Either way, that peak voltage in a fully charged cell still stands out there, and overvoltage damage in semiconductors is generally lightning fast and final. Those 10V rated charge pumps are more than likely the same manufacture as 12V rated ones that just didn't consistently survive 12 volts at the bench but could be safely rated at 10. When teaching electronics it is an often shared rule of thumb to try and overspec your parts 50% or more, to extend their lifetime; we generally see 35V electrolytic caps instead of 25V ones on a 24V rail even if that rail was regulated within 100mV.

Apologies to Koala and the rest, I believe we should move this to General Electronics under nickel battery behavior ... if we really need to keep dragging this on, KT.