Battery Vehicle Society

I've redrawn the schematic using Kicad. It's been a while since I last tried it, so far so good. Components have been moved on the schematic just to try and make it clearer, they have no reflection on how I am thinking of positioning them on the actual pcb. The component reference system I was using had to be changed a bit. So capacitor C1 on the first cell was referenced as C1a now becomes C101, C1d is now C104, etc. I had to do this because Kicad uses the letter suffix for multi part components such as op-amps that may have 4 op-amps on one chip.

Looks ok but we said a lower value resistor for the transistor driver IIRC

We also said 10uf for the bigger cap

R3xx (driver resistor) at 330R gives a current of around 10ma @ 3.6 volts. A 220R would be 16ma @ 3.6 and 19ma @ 4.2 volts. I could change it to 220R if you think that would help. I will also update C2xx to 10u. I am working on the layout, it's a bit tricky making it all fit. I am thinking of stacking the ICSP headers into pairs of 2. This will allow programming from the edge of the board with 90 degree headers.

I've still got some work to do, but here are a couple of screen shots of the board in 3d view. I don't have a 3d model of the optos so they just show pads at the moment and the board is about half routed. I've paired up the ICSP headers so that you have the option of fitting straight or 90 deg pins, or nothing at all. I will add some holes in the crner for mounting, although I plan on sliding them into a case with grooves. Any comments?

Impressive models.

If we are going smd I think we should have smd leds, transistors and caps as well.
In fact everything except the 3w resistor which should be the only component on other side.

But if this works then we can move to full smd for the production version

Not sure how you will attach header plug to the icsp in a stack?

I was originally aiming for just the 3 watt as through hole. The transistors changed to through hole actually to save space. Using a smd transistor means you then need a via to get the signal to the opposite side of the board and was taking up more space. Same with the leds. The through hole cap is C2xx, only required if we decide to use pwm, so it is optional. C1xx is smd and again the through hole eliminates some vias.

Not sure how you will attach header plug to the icsp in a stack?

One of these cut in half. http://www.rapidonline.com/Cables-Connectors/10-10-Way-Right-Angle-Header-22-0815

retepsnikrep wrote:If we are going smd I think we should have smd leds, transistors and caps as well.
In fact everything except the 3w resistor which should be the only component on other side.

I'm still working on the pcb layout, unfortunately I haven't been able to spend all the time I would like on it. After trying several different layouts, I am going to try with everything smt except the 3w resistor and leds (but if anyone has a suggestion for a right angle led let me know).

For the transistor this looks like a smt equivalent. http://uk.farnell.com/fairchild-semiconductor/bc818-40/transistor-npn-sot-23/dp/9846662

I'll change C20x to a ceramic 0805 part and the ICSP pads may need to change to smt pads, I'll see how I get on.

I think I've cracked it. Here's what I've got.

Greg

Looks nice but for auto assembly in china can the machines cope with parts at an angle I've only ever seen pcb's with stuff at 90 degrees or aligned vertically or horizontally?

Also having the pic on underside oposite resistor might cause it to drift as it heats up

Peter

retepsnikrep wrote:Looks nice but for auto assembly in china can the machines cope with parts at an angle I've only ever seen pcb's with stuff at 90 degrees or aligned vertically or horizontally?

I have thought this could be a problem, we will need to find the answer to this question. If it does increase costs then probably the easiest solution is to rotate all smd components 45 degrees, then the board can be loaded at a 45 degree angle and hey presto, all compents are 90 degrees.

Also having the pic on underside oposite resistor might cause it to drift as it heats up

I did try about a half dozen different approaches to get the resistor away from the pic. Now I'm not saying it can't be done, but I haven't been able to get it to work. I can post the Kicad files if someone else wants to have a go.

I think we can do a couple of things to minimize the pic heating up.

First, boards must be mounted with the resistor up or the board vertical. Heat rises so don't orient the boards with the resistor below the pic.

Second, some sort of heat insulator between resistor and board. It must be non conductive and heat resistance, maybe some bare FR4 sheets?

Third, we can do some testing and see how much drift we get with increased heat.

Fourth, don't balance on a regular basis. This last idea comes from talking to Evan Tuer a couple of weeks ago about his 106 electric which he converted from the original nicad to lithium a couple of years ago. He has come to the conclusion that he will balance the pack once a year. He has stopped doing regular balancing on his pack, finds it unnecessary. Of course your mileage will vary and maybe he has got lucky with his cells.