creating phase delay

[jimmylovesni]

anyone have an idea on how to create a phase delay which is variable. Im not talking the start 180 degree flip flops though, but more like 10-20 degrees. I know this is not that hard to do with a u controller but need to try with standard ICs. A 555 comes to mind but I havent been able to get a good solution working.

 

[VEBill]

Variable TIME delays are typically quite simple. So if you are dealing with a fixed period, then creating a variable phase delay is equally simple. On the other hand, if the period varies then it becomes more complicated to create a relative phase delay.

Many real world applications would increase the number of pulses per cycle. If you tried to make do with only one pulse per cycle and tried to create a phase delay - and then the period was suddenly accelerated - then there's no method to be accurate in real time. E.g. Radar antenna typically use one Azimuth Reference Pulse (ARP) and a large number of Azimuth Clock Pulses (ACP) per rotation.

Doing all this with basic gates would be a time sink. There are specialist chips for some of this, perhaps in the automotive ignition field?

 

[MacGyverS2000]

Knowing what the application is would help us provide more useful answers... a variable phase delay is something most only care about when it comes to equalizing comm signals (and similar).

Dan - Owner

http://www.Hi-TecDesigns.com

 

[jimmylovesni]

im trying to create a pump from three piezo disks. The have to fire just micro-millliseconds from each other to get the timing for the pumping action.

I was thinking a astable 555 but was unable to get it to work. It seems to always start with the caps charged so there is no phase delay from the rising edge. vice versa, in monostable mode, it just fires on the falling edge.

Any ideas on the time or phase delay?
thanks

 

[MacGyverS2000]

Time delay and phase delay are two completely different things, hence my question. It sounds like you want time delay (being an Automotive guy), as VE1 started with, but your mention of degrees leads us down the path of phase delay.

Do you need microsecond to millisecond delays, or do you need nanosecond ("micro-milli"?) delays... a 555 isn't going to get you nanosecond. How accurate/repeatable do the delays need to be? Consider first you're usually working with low-tolerance components with a 555 (caps that are 10%, resistors at 1% before the money starts to increase, etc.), so tight timing is a complete crapshoot.

Why the insistence of using a 555 when a micro will do this without blinking an eye?

Dan - Owner

http://www.Hi-TecDesigns.com

 

[jimmylovesni]

ive tried the 555 without any luck, thats the only reason i mention it.

i think the timing delay and phase delay are somewhat similar if the timing delay can be triggered from a control pulse.. in that manner, it is always a fixed difference from the rising edge. Im not sure how to do this anyway.

I am looking for approximatly microsecond difference in phase delay with a frequency of around 1khz

thanks

 

[MacGyverS2000]

Stop using one term if you mean the other... you expect a time delay of 1 us. I'd use a $2 micro and be done with it.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[itsmoked]

Perhaps this recent EDN article?

It's the forth one down.

http://www.edn.com/file/26015-Programmable_driver_targets_piezoelectric_actuators_PDF.pdf

Keith Cress
kcress -

http://www.flaminsystems.com

 

[jimmylovesni]

that article is a microcontroller based application, it doent really have the phase control that we looking for, that is 10-20 or 30 degrees.

i still dont see that a time delay can be used to simulate a change in phase, its periodic with respect to the driving pulse

yes, a microcontroller, i know its pretty simple with it

 

[MacGyverS2000]

A phase delay assumes you know the period of the driving signal... you have mentioned no such spec. You have only mentioned a time after a start signal between which the multiple elements must "fire". The elements could be driven at 5 Hz, 50kHz, 5GHz, etc. but activate 1us apart from each other. There is a major difference, and continuing to use one term when the other is meant serves only to confuse the issue. The latter is a time delay that has nothing to do with any phase difference between the elements. If you need the elements driven from the same drive signal where each is a certain number of degrees of phase off from the other, then specifying a time difference makes no sense (unless we are to infer the drive frequency from such details, which is a questionable tactic).

I still fail to see why you won't use a microcontroller, particularly since you have admitted it would be easy to do so with one.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[OperaHouse]

It sounds like you could just use some cmos gates with a RC time delay several times to create the delays for a charge pump. Maintaining a waveform is not necessary. There are also those bucket brigade chips used for audio delays. If frequency changes those phase degrees may not be important. It is not clear to us what you need. Maintaining a phase relationship may be making it overly complex.

 

[pstuckey]

I think you can probably find a chip here that will do what you need.

http://www.datadelay.com/index.htm

Peter

 

[jimmylovesni]

thanks all