Howard, many thanks for all this information. I have been trying to make something similar for a while now, but with no access to AMB equipment (I am not interested in compatibility).

I have been playing with your transponder circuit this weekend, right now just the LC tank portion with a 47p, 62p and a square wound core using magnet wire which is slightly bigger than the one in your first post. Right now I'm feeding a square wave at 5VDC from a function generator, through two 100R resistors as if it were driven by the XOR gates.

Sweeping the tank shows it is resonant at around 5.22Mhz (I will adjust capacitors later). I measured directly across the inductor and got about ~13Vp-p.

One thing confuses me though, I am trying to pick this up with two loops of wire. This is fed straight into the scope.

One is a large square of thin stranded wire, roughly 30" x 12". I get a mostly clean 5.22Mhz sine wave from this, even at a fair distance.

The other is an 18AWG loop at 18" long. I get practically no 5.22Mhz, even right next to the loop. If I adjust the frequency to 12.2Mhz I get a sine wave at the loop, but scoping the tank itself is very messy. It isn't a resonant frequency of the tank. I am guessing here I have built a poorly coupled transformer!

I did not think that reception should depend on the size of the loop, have I missed something? Or is it just because my shorter loop so small that it stands no good chance of receiving the 5Mhz signal?

I can't explain at the moment why you are getting such disparate results from your two receiving loops, but I can offer some advice that will help you with your experiments:

First, the transponder tank was designed for differential (bridge) drive, but your signal generator very likely has a single-ended output with 50 ohms source impedance. Let's refer to the transponder schematic:

http://www.rctech.net/forum/11809951-post112.html

For single-ended drive, eliminate R5, connect the signal generator ground directly to the junction of CT1 and L1, and drive R4 with the "hot" lead from the signal generator. Also, R4 would now need to be 200 ohms (R4+R5), but the generator already provides 50 ohms, so make R4 150 ohms.

Don't make any scope connections directly to the tank, as this would load the circuit unnecessarily. Use a single loop of wire, about 2 to 3 inches in diameter, connected to the scope probe to "sniff" the magnetic field of the tank. (I used the ground clip of the probe as the loop.) Loose coupling is exactly what you want! The scope probe should be set for 1M or 10M impedance. Center the tank circuit in the pickup loop.

Larger loops for use on the track will benefit from being tuned to 5 MHz, at a relatively low Q, as shown in the loop amplifer schematic from my decoder thread, here:

http://www.rctech.net/forum/12118142-post206.html

It makes no sense to feed any noise or extraneous signals to the electronics when they are easily filtered out first!

As an aside, the best design for a differentially-driven transponder tank would be totally symmetrical; that is, CT2 would be split into two capacitors of double the original value, with each one driving one end of L1. I didn't do this due to board space limitations, but if the layout were done using SMD components like 0402 capacitors, then I would go this route (perhaps more for aesthetics than any measurable difference in operation).

Thanks for your reply - apologies it took a while to respond but I haven't had much time to experiment. Attached are some pictures as I can't include them yet.

The transponder is hanging from a piece of string 18 inches from the table, where a large pickup loop of stranded wire is fixed. A scope probe (with GND lead shorted to tip) is hung around the inductor.

The yellow trace is off the large pickup loop, fed through a simple RC filter. The blue trace is from the probe. I don't get quite as high Vpp from a distance but it seems fairly clean - and 18 inches is much further than I'd hope for.

Right now I'm just running some C code on the PIC to toggle periodically, I haven't tried to match it to the required 4 cycles yet. I'm using a 12F1840 with 4x PLL though so this wont be hard.

You can see that I have switched to using a slightly different inductor, I wrapped a number of turns around a ferrite rod. This measures around 3.5uH using an LCR meter and because of this my capacitors are much larger than yours. I tried this as I'm hoping to use it on karts rather than RC cars and this seems to be how the kart transponders work - I assume the magnetic permeability of the ferrite material helps it to propagate further. I know that the AMB kart transponders operate on a different frequency but as I said previously I am not worried about compatibility.

Does the scope screenshot look roughly correct to you? My next step is to build your loop amplifier and see if I can get it to work nicely.

If I'm not mistaken the old AMB20 system used the battery as the core for the inductor,.. not sure if any other systems used this or not.

The waveforms look very nice! I might suggest using a little lower Q so the waveform "rings up" a bit quicker, but it's probably not necessary.

The ferrite core increases the effective aperture of the transponder's "antenna" loop, and of course increases the inductance. I don't think it will be necessary for what you are doing, but if it intrigues you, go for it! Also, let us know if if you discover that there are other advantages to it as you experiment further.

If you don't need compatibility with AMB, then life will be much easier for you, as there won't be any need to replicate data trains from existing transponders.

Now that would be ingenious! I love designs that exploit every use for every component!

Well it took a few hours this afternoon but here's my PCB. A little larger than some of the others on here, but I wanted to build in a linear regulator. It would be possible to use the same pads for a diode regulator if preferred.

The only thing I could not easily fit on a single side was a resistor divider for input voltage monitoring (this could be encoded in the transponder output, like I presume the AMB ones do). I'd prefer to keep it on one side for now and add features later if required.

The only different things are a status LED, two mounting holes and strain relief hole for the input cable.

The ICD header is directly compatible with the PICKit3.

Thanks again to Howard for all his work on this, I'm looking forward to building some boards.

That's very nice work!

I'm concerned that your inductor has very few turns compared to the original design. That will give a much lower output unless you have increased the drive current to compensate for it.

Very good point, I will add some more turns before ordering. I looked at both your design and the one from Gerrie and ended up in the middle. Gerrie has four layers though, so it would probably be best to get to at least 10 turns.

I was confused at first by the best way to use both sides, I assumed three things:

- crossing the tracks on different layers does not matter (e.g. my top left)
- the trace should always continue in the same direction
- there is a single trace with one beginning and end, rather than "doubling up" the trace on different layers

I hope these are correct.

What were the dimensions of your final PCB?

Your trace arrangement is fine. The traces are two spirals in series, joined to each other at the outer edges with a via. The current in each spiral flows in the same direction so the fields reinforce. On the original layout, the spiral on the top layer goes counter-clockwise from inside to outside, and the spiral on the bottom layer goes counter-clockwise from outside to inside. (This can also be clockwise on both, as on your layout.)

Using more than two layers requires fewer turns on each layer, while making the X-Y dimensions smaller requires more turns.

You can use the original PC layout as a starting point for a new layout, especially if you are using FreePCB. The original layout is 1.25" square, the same as yours.

I modified my board today to have 11 turns on both sides. It is a tiny fraction bigger than before but identical otherwise. Thanks for your advice.

Next step is to get the decoder working!

A great work of development HowardCano
their work is followed from spain with a strong interest
Congratulations and sorry for my bad English

Thank you for your support!

Well I have completed some code for the transponders now. I can transmit any identifier, plus battery status and other data (e.g. a CRC).

The actual code which sends the modulated data is a tight ASM loop of 7 instructions. Finding the right combination to keep within 4 cycles (I am running the PIC at 20Mhz) was good fun. This does mean I can vary the modulated data in any way though.

Unfortunately I am still waiting for my PCBs, so I'll be going back to the decoder for a short time.

Thanks again Howard, this has been keeping me busy and entertained!

Is someone sells PCB transponder ?
PM please.

Thanks.