Thanks RJM...
I built one of those before but using an op amp for the amplifier. I didn't think it worked very well.
But from the National Semiconductor 1980 Audio/Radio manual page 2-62 it states:
Unlike semiconductor noise sources, the MM5837 provides very uniform noise quality and output amplitude.
The MM5837 uses a clocked shift register to generate noise which I thought was a much better approach.
I built one of those before but using an op amp for the amplifier. I didn't think it worked very well.
But from the National Semiconductor 1980 Audio/Radio manual page 2-62 it states:
Unlike semiconductor noise sources, the MM5837 provides very uniform noise quality and output amplitude.
The MM5837 uses a clocked shift register to generate noise which I thought was a much better approach.
Hi,
I reverse engineered the internal circuit of I think the MM5837 and then built a version from shift registers (from memory I think that it's 17 stages) and Exclusive OR gates for generating a psuedo random digital signal. Used for testing Motorway 'Information' sign communication signals. The signal appeared pretty random on the valve oscilloscope that I used. I guess that up to the clock frequency it is a good white noise source.
I reverse engineered the internal circuit of I think the MM5837 and then built a version from shift registers (from memory I think that it's 17 stages) and Exclusive OR gates for generating a psuedo random digital signal. Used for testing Motorway 'Information' sign communication signals. The signal appeared pretty random on the valve oscilloscope that I used. I guess that up to the clock frequency it is a good white noise source.
At one time I was looking for highest noise volts out for minimum microamps in and I wound up with 1M into a revered biased 3904 base emitter junction feeding a mu follower. Drawback is you need at least 10V supply to make it practical. More than a 9V battery except brand spanking new, anyway.
Last edited:
Temperature compensation is left as an exercise to the reader....!
Pretty weak variation around room temperature, though in the UK it is probably noticably lower.
The chip's output is "uniform" because it repeats every couple of seconds. If you measure its peak output for a few seconds, you've measured the highest peak output it will ever give. It's like a two-second tape loop of recorded noise that gets played back over and over (once you hear it loop through a few times it actually SOUNDS like a two-second tape loop of recorded noise). This can be useful when you want to make the noise drive something to just below clipping, and need to know that it will never clip. With an analog noise source it may eventually clip, as the level is "truly" random and you don't know what the peak may be.
Here's a description of a PIC microcontroller replacement:
White Noise Source based on PIC 12F processor
The PIC code described is "better" in that it has a much longer period, but once gain that's not always what you want.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.