Field testing modified MPP

[CAS]

Thank CAS. Basically the first sample needs to be taken after the flyback reaches flat line. I'll post a pic shortly of what I'm talking about.

The earliest I was able to sample was at about 7us. Thats when the flyback leveled out. I had to change the damping resistor a few times to get optimum results over several coils. Best practical results were when sampling was about 9 or 10us.

[mickstv]

This was just using an old NE5532, the first stage has a couple of 1n4148's added in parallel between pin's 1 and 2. The diodes speed up the decay time by about 1us with this circuit.

[mickstv]

The earliest I was able to sample was at about 7us. Thats when the flyback leveled out. I had to change the damping resistor a few times to get optimum results over several coils. Best practical results were when sampling was about 9 or 10us.

You'll find that minimum sample will increase when the current limit mod is removed.

The picture in my previous post was with a ML mono coil, and the TX time was about 120us.

[CAS]

This was just using an old NE5532, the first stage has a couple of 1n4148's added in parallel between pin's 1 and 2. The diodes speed up the decay time by about 1us with this circuit.

Connected between output and inverting input? I had them in a couple of test preamps I was working on using a LME49990. Didnt seem to make a huge difference on the output of the preamp.

[mickstv]

Yeah between negative and the output pin. I have a feeling any improvement will only be seen in lower speed op-amps like the NE5532 or NE5534. The LME49990 is a high speed device, so the difference between diodes or no diodes wont be much.

A LME49990 would be a waste of time in any PI that uses a 1k resistor and protection diodes in the frontend. The noise from a 1k resistor will be higher than the noise level of the LME49990.

[Jose]

This was just using an old NE5532, the first stage has a couple of 1n4148's added in parallel between pin's 1 and 2. The diodes speed up the decay time by about 1us with this circuit.

Mickstv
The mod, which you mentioned, comprises two diodes in total, or four, as the outline of "Cas" ..
Thank You
Jose

[mickstv]

Mickstv
The mod, which you mentioned, comprises two diodes in total, or four, as the outline of "Cas" ..
Thank You
Jose

Hi Jose, I was just using two diodes. When looking at the output of the first opamp, the two diodes limit the output to +/- 1v max, whereas four diodes increases the limit to +/- 2v max. The decay is the same for either two or four diodes.

[6666]

Hi Cas
I know we have been asking you lots of questions and asking you to do things, but if you dont mind too much
would it be possible to just hand draw Davor's pulse train please.
I have been trying to nut out your code, and I think I have it, but there is a couple of bits of code I dont understand
the reason is, I'm interested in comparing another circuit , thanks.

[CAS]

Hi Cas
I know we have been asking you lots of questions and asking you to do things, but if you dont mind too much
would it be possible to just hand draw Davor's pulse train please.
I have been trying to nut out your code, and I think I have it, but there is a couple of bits of code I dont understand
the reason is, I'm interested in comparing another circuit , thanks.

I have deciphered my notes and this is what I had as the base calculations:
MAIN DELAY = x (set between 7us and 25us)
MAIN SAMPLE = y (again various but between 10 and 60us - If it got too large the calculation would go into error as total would be larger than TX PERIOD )
GB DELAY = z (I used various delays of around 25 - 35 us)
GB_SAMPLE_WIDTH = ((MAIN_SAMPLE_DELAY + MAIN_SAMPLE_WIDTH + GB_DELAY)*(MAIN_SAMPLE_WIDTH / MAIN_SAMPLE_DELAY))
EF_SAMPLE_DELAY = TX_FREQ - TX_WIDTH - MAIN_SAMPLE_WIDTH - (2 * GB_SAMPLE_ADJUSTED) - EF_SAMPLE_WIDTH - 10 the 10us is just a little fudge factor to give small buffer between end of EF SAMPLE and next TX PULSE
EF SAMPLE = GB SAMPLE - MAIN SAMPLE

Its hard to draw out pulse train as I was using a pot (pot2) to vary MAIN SAMPLE DELAY
I also used pot 1 to allow a small fudge factor in GB calculations. I didnt want to get out in field and find my code was wrong and it didnt work. So what I did was have a adjusted GB sample width - subtracted 50 then added what ever pot value was. that gave me a value of +/- 50. With pot in center the 50 was added, cancelling out the subtracted 50

If you look at this thread http://www.geotech1.com/forums/showt...Ground-Balance you may be able to tie the code to Davors description. Ignore my initial timings on that page as they were way off.

[6666]

Hi Cas
I'm a bit confused and no code expert, please correct me if I'm wrong
I've been looking through the text comments in your code, and you seemed to have labeled output RD7 as pin 36 a number of times

LATDSET = (1<<7); // SET PIN36 or PORT D7 high

[CAS]

Sorry Pin 36 is D6 - I made so many changes over the weeks I often didnt correct the comments.

pinMode(34, OUTPUT);//RD5 TX_PULSE
pinMode(36, OUTPUT);//RD6 RX_GATE_1
pinMode(37, OUTPUT);//RD7 RX_GATE_2

I never really intended anyone else to have to try and read my code. The comments were what I made when I first started the project and I forgot to update them as I went along.

[CAS]

6666 I see you are located in Melbourne - If you are ever down Geelong way send me a message as you are quite welcome to drop in to the workshop to have a chat about the circuits/code. The workshop is located in Breakwater, just a few minutes from the Geelong CBD.

[6666]

6666 I see you are located in Melbourne - If you are ever down Geelong way send me a message as you are quite welcome to drop in to the workshop to have a chat about the circuits/code. The workshop is located in Breakwater, just a few minutes from the Geelong CBD.

Hi Cas , re the code thanks for the reply, I only noticed the text comments and that something didnt look right because I was plotting the pulse train and all the action was happening on pin 36, so I looked up a reference book that explained the LATDSET command and worked back to the correct pins from there, but its all good at least I learnt something.
Thanks for the invite to visit, I'm way over on the other side of the bay in the eastern suburbs, and rarely get to your part of the world but maybe one day I could be over that way, if so will send a message , cheers.

[CAS]

One other thing I just noticed is that the TX period is now 800us not 1000. Its another thing I forgot to change in the comments. I read somewhere about a calculation from mains frequency to help reduce EMF. Not sure if it helped or not and I cant find where I read it.

[Chet]

Hi Cas,

Thank you for sharing your code and modifications. I found the information below from one of Carl's Posts. I think it could be worthwhile to add another differential integrator channel as he describes.

"When we implement subtractive GB we normally sample shortly after the target, amplify the GB sample, and subtract from the target sample. This is usually done with a completely separate channel and its gain is the GB control. Again, matching the taus is critically important, and if the target sample has an Earth-field subtraction then so must the GB sample:


BasicPIGB.jpg
Timing is now:
BasicPIGBTiming.jpg"

From http://www.geotech1.com/forums/showt...957#post193957

Have a look at the above .jpg files at that location.

Have a great day,
Chet