Whites TDi Timing help

[Sean_Goddard]

I have both the ACTUAL Whites TDi schematic, AND the TDi "Proto" schematic from here (NO I WONT give out any copies, I am sworn to keep it confidential, so DON'T ASK) but I need to know the following;

The pulse delay is adjustable yes?

OK, if it is, then do the other signals (S2, 3, 4, & 5) TRACK the initial S1 sample pulse, OR is the "main" sample pulse independent?

S2 is 34us from the end of the Tx pulse (and 9us from the end of S1) IF delay is set to 10us, SO, if I set the S1 to 5us, does S2 still occur at 34us, or does it similarly occur a fixed time after S1 goes low?

ALSO, to vary the overall Tx frequency, does the delay between S4 end , and the next Tx pulse vary, or do ALL the pulses compress (timewise) that is to say, do ALL pulses retain the same PERCENTAGE position of the overall Tx cycle (if that makes sense)?

What I plan on doing, is programming a PIC micro using two of the timers, One timer will be dedicated to the S1 sample, the other to generating S2 - S4 in the following manner;

Main PIC clock = 40MHz (25ns). This will also drive the timers.

15us sample pulse width = 600 clock cycles

Set an array of;

1. 65535 - 400 (10us) - Main Sample delay
2. 65535 -600 (15us) - Pulse Width
3. 65535 -360 (9us - Derived from 34 - (10+15)) - Delay S1end to S2)
4. 65535 -600 (15us) - S2
5. 65535 -3440 (86us) - S2end to S3
6. 65535 -600 (15us) - Pulse Width
7. 65535 -600 (15us) - S3end to S4
8. 65535 -600 (15us) - Pulse Width
9. 65535 -600 (15us) - Delay before Tx signal starts again

After the Tx pulse, each value will be loaded in turn, from the array into the timer where it will count UP and generate an interrupt at each "rollover". This array "INDEX" is the "Timing Stream Function Counter" which will be used to determine what should be done to one of five output pins (Tx, S1, S2, S3, & S4).

This is FAR more accurate than the Tx, delay, pulse, delay, pulse, delay, pulse, delay, pulse used in the HammerHead (Sorry George) as you should NEVER use "software delays" within an interrupt routine, far too much chance for the stack to fall over if a glitch occurs.

[Carl-NC]

Pulse delay is adjustable. S2 moves with S1 so there is a fixed delay between them. As I recall, the delay from S2 to S3 is fixed. Don't know what S5 is. It would be better to have the S2-S3 delay vary with frequency as that will vary the comb filter response and may do a better job of EMI suppression.

If you use an 8 bit PIC then the timers run at fc/4, or 10MHz in your case.

[Sean_Goddard]

Cheers Carl. Much appreciated.

[Alexismex]

Hello Sean,
"(Sorry George) as you should NEVER use "software delays" within an interrupt routine, far too much chance for the stack to fall over if a glitch occurs".
I wish you, good,good luck in keeping the noise generated by the PIC at a very low level, to maintain the same performance of the analog circuit of the TDI !!!

[Alexismex]

I have both the ACTUAL Whites TDi schematic, AND the TDi "Proto" schematic......and for the schematic , you don't need to show off at the forum...the time of human life is not so relative!

[Sean_Goddard]

I wasn't showing off, I was making a point that when I told one person I had the Minelab Sovereign schematics, I was bombarded with "Can I have a copy"? Not just from this forum.

I'm putting the micro in a screening can on a separate PCB. I'll bring out the sample pulses as differential, phase matched lines (slew rate limited), then convert to single ended at destination.

My "other" job is designing military radar, radio, and sonar system. I'll let you know how I get on. Might just design it then archive the files as I have with many other projects that never got built.

[Qiaozhi]

This is FAR more accurate than the Tx, delay, pulse, delay, pulse, delay, pulse, delay, pulse used in the HammerHead (Sorry George) as you should NEVER use "software delays" within an interrupt routine, far too much chance for the stack to fall over if a glitch occurs.

Hi Sean,

What's this (Sorry George) about?
Where have I used software delays in the interrupt routine? Both the Voodoo Project and the ANPI use timers to create the pulses.

[aft_72005]

I have both the ACTUAL Whites TDi schematic, AND the TDi "Proto" schematic......and for the schematic , you don't need to show off at the forum...the time of human life is not so relative!

Yes..it is obvious that our life is short..so it is better to pass on the experiences to others, otherwise it will be buried!!!!

[aft_72005]

Hello Sean,
"(Sorry George) as you should NEVER use "software delays" within an interrupt routine, far too much chance for the stack to fall over if a glitch occurs".
I wish you, good,good luck in keeping the noise generated by the PIC at a very low level, to maintain the same performance of the analog circuit of the TDI !!!

I remember that somewhere in this form... there was a whites TDI schematic... but now I can't find it by searching... I am also interested in reading its circuit diagram if someone has it and upload it again.

[aft_72005]

I have both the ACTUAL Whites TDi schematic, AND the TDi "Proto" schematic from here (NO I WONT give out any copies, I am sworn to keep it confidential, so DON'T ASK) but I need to know the following;

The pulse delay is adjustable yes?

OK, if it is, then do the other signals (S2, 3, 4, & 5) TRACK the initial S1 sample pulse, OR is the "main" sample pulse independent?

S2 is 34us from the end of the Tx pulse (and 9us from the end of S1) IF delay is set to 10us, SO, if I set the S1 to 5us, does S2 still occur at 34us, or does it similarly occur a fixed time after S1 goes low?

ALSO, to vary the overall Tx frequency, does the delay between S4 end , and the next Tx pulse vary, or do ALL the pulses compress (timewise) that is to say, do ALL pulses retain the same PERCENTAGE position of the overall Tx cycle (if that makes sense)?

What I plan on doing, is programming a PIC micro using two of the timers, One timer will be dedicated to the S1 sample, the other to generating S2 - S4 in the following manner;

Main PIC clock = 40MHz (25ns). This will also drive the timers.

15us sample pulse width = 600 clock cycles

Set an array of;

1. 65535 - 400 (10us) - Main Sample delay
2. 65535 -600 (15us) - Pulse Width
3. 65535 -360 (9us - Derived from 34 - (10+15)) - Delay S1end to S2)
4. 65535 -600 (15us) - S2
5. 65535 -3440 (86us) - S2end to S3
6. 65535 -600 (15us) - Pulse Width
7. 65535 -600 (15us) - S3end to S4
8. 65535 -600 (15us) - Pulse Width
9. 65535 -600 (15us) - Delay before Tx signal starts again

After the Tx pulse, each value will be loaded in turn, from the array into the timer where it will count UP and generate an interrupt at each "rollover". This array "INDEX" is the "Timing Stream Function Counter" which will be used to determine what should be done to one of five output pins (Tx, S1, S2, S3, & S4).

This is FAR more accurate than the Tx, delay, pulse, delay, pulse, delay, pulse, delay, pulse used in the HammerHead (Sorry George) as you should NEVER use "software delays" within an interrupt routine, far too much chance for the stack to fall over if a glitch occurs.

I have this .... may be help you

[Sean_Goddard]

George, the Timer1Delay(Delay) you used DOES, indeed use the timer, but if you work down into the code produced by the compiler, it uses a interrupt to "work" the delay. This is using an interrupt within an interrupt and I was taught this should not be done. I accept that coding practices may have changed.

The danger of doing this is that you are nesting interrupts and the stack values. Whilst in this case, you are probably OK to do this, in a "real world system" you run the risk that if an interrupt occurs whilst executing an ISR, the stack does not truly represent the true return address, thus it can fill up and cause an overflow.

In the application you have written, it is OK as there are no "external" or unexpected interrupts, but in general I, for one, would never use this method which is why I asked (quite a while ago) how you have done the delays.

Please note, this is NOT a criticism, just an observation. I would NEVER presume to tell you, or many of the members of this forum what to do, I'm not smart anough for that