PI CHIP

[moodz]

Hi Moodz,

I don't know how you have set up the pulse train, but if its using the output compare module, you only need 1 oc for each channel. After the + sample has been taken, re-assign the OC to the - sample pin....

Cheers Mick

....heh heh ... You are right Mick ... It's using all the OCs .. I just implemented the timings as per the diagram .. With pulse steering I will have all the sigs and one OC left over for ... Channel 3 Or even the AGB. I only hammered the basic code out to test for any jitter or funny business in the pulse hardware that microchip implements. So far so good.
Now to add the QED timings from the patent ...for experimental purposes only.

[Mechanic]

Hi Moodz,

I don't know how you have set up the pulse train, but if its using the output compare module, you only need 1 oc for each channel. After the + sample has been taken, re-assign the OC to the - sample pin....

Cheers Mick

[moodz]

The pix are posted at a well known unmentionable web site under this topic.

"pulse generator for doing detector stuff"

...third try ...

comparison of the TX pin to the other pins ... the pulse train is similiar to an expired patent and old detector.

[moodz]

The pix are posted at a well known unmentionable web site under this topic.

"pulse generator for doing detector stuff"

[moodz]

...try to upload pix again ...

damn ... cant upload anything even a screenshot of the upload error ?? doh.

[moodz]

Hi I have to agree with Stefan his method does make the boards easier to drill.

Moodz OK after more testing at various timings I find that your code does not allow for sampling in the on time therefore not suitable for the GG Tinkerer project will the UNO32 code allow for it?

Regards, Ian.

It might ... however you will need to provide the timing diagram for the TD.

In the meantime I found another timing diagram lying around on the web ( for SIDO ) and popped some code into the UNO32 to see how it would do .. indeed it is not bad ... very low CPU overhead around 5%, arbitrary pulse trains with patterns that are only limited by the RAM to store them.....and the timing clock is 80 MHz so thats 12.5 nanoseconds of resolution on adjustments. Each mark and space in the pulse train is fully adjustable on the fly and its all generated with 50 lines of code ... not bad if I say so myself

There are 5 outputs on the UNO32 implemented so far ... Pin3 = TX Pin 5 = channel 2 sample A Pin 6 = channel 2 sample B Pin 9 = channel 1 sample A Pin 10 = channel 1 sample B. The ground balance samples are another story ... I ran out of pulse hardware on the PIC32 .. but doable. ... but who needs ground balance when you have the patent pending "moodz" AGB system ....right ? ...

This system would be very busy to manually enter a pulse train and adjust each pulse timing on a simple LCD display ... so I am thinking a bluetooth interface to program it from a smartphone ( prob android ).

Heres the hex so you can verify ... note the menu does not work !!!! ... this is a UNIPI hack to get the pulse generator only going ... its not adjustable ...

I would upload the pix and the hex .. but nothing is uploading to geotech at the moment

[IBGold]

Hi I have to agree with Stefan his method does make the boards easier to drill.

Moodz OK after more testing at various timings I find that your code does not allow for sampling in the on time therefore not suitable for the GG Tinkerer project will the UNO32 code allow for it?

Regards, Ian.

[Stefan]

Hi,

I have a small suggestion with regards to the hole sizes in PCB for home construction. I always reduce the size of the pad holes to minimium size. This allows the pad to be full size after etching and also the small hole size guides the drill to the centre of the pad during drilling. A simple idea but you will find it works well. Nothing worse than drilling out the side of a pad and breaking a connection.

Regards,

Stefan

[sido]

It could be a bug but more likely the original code was never designed to handle very short times ... More conventional timings like 100 us on and 400 off etc. there is prob a race condition in an interrupt handler.

Hey Moodz, i assume you mean T(race) Heheheheh

Quote from Moodz.........
"So you just want an arbitrary pulse sequence to be fired to then repeat. No real prob with that. I'd rather use an FPGA to do this ... Actually I did use one ..but that's another story. So the short answer is yes ... We can do that ...just need to pull the finger out and code it up."

Yes....If it is not a big software code adjustment to add for each pulse output of UNIPI V300 and providing the chip can handle it, I would appreciate it, and assume other's here will too.

We may have to call it UUNIPI (Ultimate UNIPI)....or what ever you may think is suitable for its purpose, maybe UNIPI V400.

BTW, if you say you have already tried it, what do you mean?......If it's something recent, no need to answer...Dont want to know.

I would prefer the UNIPI V300 to add any of the above modification if possible, as i see it as Ultimate uniform platform for most pulse timing needs, more than enough timing requirements that one may need for experiments.

Much appreciated in whatever you may choose there Paul. If it cannot be done with the current hardware, just let me know and i will look into the current FPGA format of your creation, if it is capable of providing the above requirements.

Though i would prefer the added timing adjustments ATM to be done with the current UNIPI, as i have been working on a PCB design based around it.

Only if you can adjust the finger as you say......versus the limited time you may have available.

Much appreciated in what you may come up with.

Cheers Sid

[IBGold]

You are over my head I no understand ring me when you get time.

Regards, Ian.

[moodz]

So you just want an arbitrary pulse sequence to be fired to then repeat. No real prob with that. I'd rather use an FPGA to do this ... Actually I did use one ..but that's another story. So the short answer is yes ... We can do that ...just need to pull the finger out and code it up.

[moodz]

It could be a bug but more likely the original code was never designed to handle very short times ... More conventional timings like 100 us on and 400 off etc. there is prob a race condition in an interrupt handler.

[IBGold]

Hi Moodz,

While waiting for the UNO32 information I reverted back to the UNI-PI to do some testing with the RX now the TX bugs are worked out I programmed the TX for an on time of 80us and an off time of 12us which was fine then I set S1 to 85us on 90us off and S2 at167us on and 172 off while looking at the outputs on the CRO TX is fine but S1 shows just before TX switch off instead of starting at 85us and I get no S2 I take it that all the timing settings should be from TX switch on, If I advance the S1 switch on and off times to try and get it in the off time I get to 92us on and any semblance of S1 disappears any ideas.

Regards, Ian.

[sido]

Thanks Sido ... question .... what interpulse tx timing max period would you like on multipulse TX ? Darn ... I should have read closer .... the SD2000 timings.

Hi Moodz, I gather i may have not phrazed it clear

What i am trying to say is that the current adjustable TX Pulse output's of UNIPI V300 can only produce a single pulse at a user selectable time (A one shot pulse of user adjustable variable width)

If it is possible, keeping the current pulse outputs setup as they are........BUT include another menu where user can select and adjust multiple complex pulses over any given time (All user adjustable).

SD2000 timings is an example of what i mean by being user adjustable.........where all the timings in that time period (1440usec in this example).........if possible can it be user adjustable as a new menu option with the existing current one.

Say i want to create a 2000usec TX Pulse Train, but say i want s1 (SEE SD2000 timing example below) to start off with 300usec TXon instead of 240usec, then followed by 300usec TXoff, then followed by 100usec TXon 100usec off etc etc etc (new timings).......This all being a user adjustable format by a new menu added and obviously some new hardcore code, providing their is no limitations to implement this in the current hardware used. (dispic4011)

Actually you wont need the current generic code for the 7 Pulsed outputs if the above can be done because if a user wants a single output pulse at any given time, this can be done so because it is all user adjustable.......and if user wants to implement a complex pulse, again its user adjustable for any give length of time.

Basically you go to menu, select s1, then enter say your first on/off pulse, then second on/off pulse, then third on/off pulse etc etc etc....where each pulse width is use adjustable also....Then when satisfied with the pulse timings you save it and the program continiously loops that new user adjusted Pulse train.

To sum it up, its modifying the 7 pulsed output program codes to allow more complex user adjustable pulses of varying degree.

I hope i explained it a little better now

Hopefully if it can be done without taking up to much of your time.
Much appreciated

Cheers Sid

[moodz]

Thanks Sido ... question .... what interpulse tx timing max period would you like on multipulse TX ? Darn ... I should have read closer .... the SD2000 timings.