high speed counters
BebopALot
Posts: 79
Hi folks, I am currently engaged in a project in which I need 24 bit counters - well actually, not the whole 24, but 19 bits of it. I want to count at 75 MHz. I initially wanted to use dedicated counter ICs for this, but have had difficulty in finding binary counters that can operate above 70MHz. The ones that I can find aren't cascadable and leave me with either 12 or 14 bits, but the right operating frequency. The only 24bit counter I have seen is the CD4045, but its slow relative to the SX and won't run above 30 MHz. My question is this:
1. does it make sense to use the SX as a dedicated 24 bit counter at 75MHz; that is - is it accurate enough? Are cycles frequently skipped?
2. just for the sake of my own curiosity, does anyone know where to find cascadable very high speed binary counters at 75MHz + or large stage counters - e.g.20 -24 bit? I have had a lot of difficulty finding what I need.
Can anybody weigh in?
1. does it make sense to use the SX as a dedicated 24 bit counter at 75MHz; that is - is it accurate enough? Are cycles frequently skipped?
2. just for the sake of my own curiosity, does anyone know where to find cascadable very high speed binary counters at 75MHz + or large stage counters - e.g.20 -24 bit? I have had a lot of difficulty finding what I need.
Can anybody weigh in?
Comments
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·1+1=10
Post Edited (Paul Baker) : 3/18/2006 5:33:04 AM GMT
What if the 75MHz was used to clock the SX? Seems like there might be a solution or a "trick" doing something that way.
BebopALot,
Is 75MHz a fixed frequency, or is this the upper limit?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe
IC Layout Engineer
Parallax, Inc.
Given that the top end on the SX is pretty much 75 MHz, trying to create a counter that counts an incoming signal at 75 MHz seems poised somewhere between dicey and impossible. There's no way to implement a 19 bit counter without using software, and that means multiple CPU instructions, and that means your SX counter will be ruinning slower than the signal it's trying to count. There might be some tricky deal using the RTCC and one of the input pins, but it all seems a bit dodgy. I'm not saying it's impossible, just not sure if it's really the solution you want.
Thanks, PeterM
Depending on what you want to do with the count.
Could you provide more details about what your trying to accomplish ?
Bean.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"SX-Video·Module" Now available from Parallax for only $28.95
http://www.parallax.com/detail.asp?product_id=30012
"SX-Video OSD module" Now available from Parallax for only·$49.95
http://www.parallax.com/detail.asp?product_id=30015
Product web site: www.sxvm.com
"Wise men know when they're right. The wisest also·know when they're wrong."
Post Edited (Bean (Hitt Consulting)) : 3/18/2006 1:38:54 PM GMT
Thanks BBAL
These are rated for 210Mhz...
So,·how are you planning on interfacing to the SX?·· I'm a little
interested, because I'm planning on doing some high frequency
counting stuff with the SX also. (That's why I found these
counters...· ...Digikey does stock them)
-Dan
·
If I am not mistaken, it looks like you are limited to 12 bits on that component as it doesn't appear cascadable - am I correct? But a good find as far as frequency. Here is what I plan to do: I have a trigger signal that is also on a carrier wave. I need for that trigger signal to start the binary counter without the carrier wave stopping it prematurely. I will use the SX to arm then monitor the trigger circuit and "hold" the trigger if it is tripped. Since my carrier frequency is much lower than my trigger speed the SX, even if it missed several cyles at 75 or 50 MHz, will have ample time to hold the trigger before the next carrier wave stops the binary counter prematurely. In this scenario the SX will create a sort of "sample and hold" use of out of faster circuitry. The SX will be "managing" this circuit.
This way you get a 24 bit counter...·· ...see attachment.
I guess my hardware is a little simpler. I'm going to build a simple frequency counter
for·1.8Mhz to 54Mhz, with 500 Hz resolution. I plan on using this for·an automatic
antenna tuner controller for a ham radio screw driver antenna controller. (A screwdriver
antenna has an inductor that is tuned with a cordless screwdriver motor as a driver)
In my case, the hard part will be to scale the voltage (From 1Vp-p to over +300Vp-p)
into a 5V logic level...·· ...but I think I've got that worked out now.
-Dan
I have used the SX48 timers a little and they are very capable. If you could explain a little more what exactly you need, I could maybe help you out. BBAL was talking about ultrasonic frequencies, then Dan was talking about 1.8MHz to 54MHz ? Are these two seperate projects ?
Bean.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"SX-Video·Module"·available from Parallax for only $28.95 http://www.parallax.com/detail.asp?product_id=30012
"SX-Video OSD module"·available from Parallax for only·$49.95 http://www.parallax.com/detail.asp?product_id=30015
Product web site: www.sxvm.com
Coming soon... Cheap 4-digit LED display with driver IC·www.hc4led.com
"Sometimes it is better to remain silent and be thought a fool, than to speak and remove all doubt."
·
I'm still testing the device at home, but am using the Toshiba chips I posted earlier
for the high speed stuff.
Bean,
I'm making a hardware frequency counter.·I need to keep checking·the frequency every
few milliseconds.·The 4040's with some bus latches make the hardware easy.
I really·need the multifunction timers for PWM outputs, so can't use them for inputs.
-Dan
·
Some while ago I needed some high speed counting done and experimented with the SX to see what it's maximium frequency was. To the best of my recollection it was in the 80 MHz range..... it's free running, and not linked to the SX clock. So making counters of any arbitrary length inside an SX should be easy.
I'll have a peek again tomorrow and post my findings.
Cheers,
Peter (pjv)
Well, I've finally had a chance to do some testing on this with an SX28 and discovered, contrary to my earlier post,·that the RTCC can be clocked from an external source·only up to half the rate of the system clock. In other words, if you are running the SX at 50 Mhz, then the RTCC can input a maximum external frequency of 25 MHz. And this maximum is also subject to the caveat that the external source must be a 50% duty cycle; an equal high/low duration.
If the duty cycle is not 50%, then to the SX it will appear that the shortest duration (either high or low) is half the input freqency, and hence it will see an apparent higher frequency. To further explain this, let me cite an example. At 50 Mhz, the period of one cycle is 20 nano seconds; 10 nano seconds high, and 10 nano seconds low. This is if it has a 50 % duty cycle. On the other hand, if the waveform were 9 nano seconds high and 11 nano seconds low, the period would still be a total of 20 nano seconds, hence 50 MHz, but the duty cycle would be 9/20, being 45%. In this case the SX would "see" the shorter of the durations, 9 nano seconds as one part of the waveform, and would "think" a whole cycle of that to be 2x9 = 18 nano seconds, or 55.55 MHz, and that is higher than its clock frequency, so it will not count correctly. This would be the case whether the 9 naoseconds were the high part, or the low part. In other words 9/20 = 45% or 11/20 = 55%. In order to be able to read at the maximum RTCC input frequency, and that is half of the processor clock frequency, then the input duty cycle is required to be 50%.
At slower input frequencies this duty cycle rule can be relaxed, to the extent as calculated by the "apparent" frequency approach explained above. So as long as the shortest duration, high or low, is not shorter than the half the duration of one processor clock period, then the RTCC input can be faithfully counted.
There is another caveat on this, and that is that an external RTCC input would usually be asynchronous to the processor's clock, and hence the certainty of the processor "seeing" an input frequency of half its own clock is not guaranteed. So the actual RTCC input frequency of a 50% duty cycle that can reliably be detected by the processor is a little less than half the processor's clock frequency.
The documentation of the SX48/52 states that "input syncing is always enabled on the RTCC input" leads one to understand why the "half clock frequency" rule exists. Eventhough the SX28 documentation does not make the same reference, by experimentation results I believe the same to be true for it.
So, the long and the short of it is that you can't count 75 Mhz into an SX...... the best you can hope for is 37.5 MHz with a 75 MHz clock, and 50 MHz with a 100 Mhz clock. Both of these I have verified on an SX 28 running at room temperature with a 5 Volt supply.
Hope this helps clear some things up.
Cheers,
Peter (pjv)
Post Edited (pjv) : 3/27/2006 7:11:15 PM GMT