My digital regulator
 

Note: I had all these parts for free. I understand this would not make sense for most people, so please don't post just to say that.

I've been having a string of bad regulator trouble. I've had them creep and over-carb my beer, leak, and otherwise just suck. Every time I looked in my kegerator, the pressure would be different. Plus, I bought a rebuild kit for my Norgren regulator and the rebuild kit was DOA! Faced with spending $70 on another annoying regulator that with my luck will probably also break, I built a digital one out of spare parts.

I supply 150PSI CO2 with one of my old paintball regulators. The keg pressure is monitored by an industrial pressure sensor and a SSR turns on a 110V solenoid valve to let CO2 in when the pressure drops below the setpoint.

I would have used a microcontroller, but I don't have any right now so I just set up a comparator circuit. It works perfect so far.

Now that this one works, of course I want to design a better one. If I designed one from scratch, I would use a microcontroller, I would use DC solenoids so that it could be battery-powered, and I would use 2 solenoids (or a 4-way) so it could bleed pressure off in case of temperature rise or a leaking main solenoid. I'm not sure where to get cheap pressure transducers, though. I have been told that Chrysler oil pressure gage sending units are 8-80ohm, 0-60psi and only cost $10 or so, but I haven't confirmed that.

Here is a schematic

You don't want to bleed pressure off if a temperature rise is the cause. Bleeding due to a malfunctioning valve would be nice, but would require temp sensing and a CO2 volumes/temp/pressure table. If you do build one using a ucontroller, that would be a useful addition for both carbing and malfunctions. Simply set the vols of CO2, and let the table set the pressure according to the temp. Line balancing may get tricky, but there is some leeway of at least a couple of psi.

Right now if you adjust the pressure down, you have to bleed a keg or drink a few pints for the pressure to actually drop. Having bleed capability would fix that.

It would also be good if there were valves that work directly off co2 pressure, then a first stage reg would not be necessary. You would use a pinhole and maybe PWM to keep pressure spikes down.

Why are you having to bleed pressure if your system/reg is working?
If it is just for when first adjusting/setting the pressure, put a PRV somewhere handy and manually vent. I could see having to mess with adjusting the pressure when going portable with a keg, but not for a keezer with kegs that have been under steady temp/pressure.

If you look at industry solutions, they have usually put quite of bit of thought into things. Most use large diaphragm regs for accuracy, as well as dual stage. A primary reg actually simplifies things compared to dealing with 1500psi everywhere. They also counteract 'decaying inlet pressure' issues. For remote control, they use 'dome loaded' regs where the pressure is set by applying a pressure 'in the dome' (usually with no feedback), and also some electrically turned screw/spring models (usually with feedback).

An electronic reg that just sets a fixed pressure seems like a complication with no benefits. A quality reg (not standard tap-rite crap) can be had for less, and less hassle. Quality gauges also help. (I put a tire/schrader fitting on my manifolds which allow me to use a single hand held gauge to set my regs with. Gauges that allow calibration would be better.)

Being able to set a specific vol of CO2 on an electronic second stage that uses a pressure sensor, temp sensor, some code and a valve is an interesting approach, though. It would simplify setting carb rates, compensate for keezer temp fluctuations, impress the ladies, etc.

You wouldn't need a bleed valve except to handle malfunctions, or for bragging rights. The extra pressure in the lines when going from higher to lower temps should disappear for the same reason the pressure is lower for the same vol of CO2 at the lower temp. You could even wrap kegs with heat film to have different temps for each beer, though that is a bit excessive. Multiple small keezers is a better approach.

You can't PWM a solenoid valve, unless you are talking sub HZ, and even then, it will drastically reduce the life of the valve. Also, you want to be able to flow CO2 at least at the rate of your pour speed for you highest carb beer.

The valves that i used to use for paintball switch in 5ms. I'm not sure about the cycle life but i assume its in the millions. Probably better to restrict flow with an orifice, though.

I routinely have to bleed my pressure down, either because i am burst carbing at 30psi or giving a fresh keg 60psi to 'seat' the lid seal.

I am interested in these 'non-taprite crap' regs of which you speak. The primary benefit of my homemade reg is that it actually regulates, ending a string of nothing but hassle that i had with commercial regs.

Do they activate in 5ms, or can they cycle continuously at 5ms periods (200 Hz)? If so, paintball may have actually provided some benefit to society after all, since I doubt that valve existed, at least cheaply, prior to being used in paintball markers. I still doubt the longevity of any common solenoid valve driven at 200 Hz, especially in this application where a single pint of beer is likely to generate ~1k cycles. Also, it would need to remain completely gas tight over its life, since even a tiny leak will drain a tank in a few days. I assume this is not a concern with paintball, since tanks are likely removed or shut off between uses.

You originally spoke of auto-bleeding pressure for spikes caused by temperature rises, which is not necessary or correct. If you are limiting it to bleeding when physically setting pressure, that was kind of my point- that the times you need to jack with the pressure, you are standing right there anyway to bleed off pressure to set the new reduced pressure by referencing the gauge. A PRV in a handy spot post-valve should suffice, although for me, kegs already have one in a handy enough spot.

I think trying to automate pressure bleed down using only hardware will be problematic. A bit of code and a processor would work better since pressure is entered rather than set visually with the gauge. Then again, using code it shouldn't require bleed down in the cases you mentioned. When burst carbing, the beer will absorb the excess pressure within hours, and unless you left a very large headspace, there is little chance that the beer would have been overcarbed without bleeding, and not with bleeding. After seating the lid, same thing- just enter the burst or 'set and forget' pressure. You could even program in seating and carbing ramps/steps.

I am not familiar with 'Norgren' brand regs, but you definitely shouldn't use paintball regs for setting keg pressures. In general, the precision and granularity of regs are directly (inversely?) related to the diameter of the diaphragm (bigger is better). The 'commercial' regs in the retail beverage industry aren't exactly precision devices- what do you expect for $30. When new, they seem to work OK for a few years. Old ones can be trouble, like you found out. Other industries have real regulators, with welding regs being the most accessible and available used (new is cost prohibitive). Just find a pressure range that provides enough granularity and precision. It takes a bit of research and hunting on craigs list or Ebay, but probably less time than building an electronic one.

Many welding regs have a combined primary/secondary, which is good for a single low pressure application, but a separate primary to drop the pressure down to at least double your highest end pressure is the preferred method for multiple reg banks. This requires low input pressure (secondary) regs to both flow enough and many times to seal properly (prevent creep). Most regs use the input pressure as the force to seal the valve/orifice based on the diameter of the plug/needle. For very low input pressures (even negative), the diaphragm is 'tied' to the needle to aid in sealing.

The valve activate in 5ms. For reference, my total programmed dwell time to trip an automag sear was 20ms. That was basically a healthy maximum without slowing down the sequence. It's possible the valve stays switched a bit longer due to inducance etc but it can't be much because the whole mechanism will cycle (actually shoot paintballs) at 20hz/50ms. I dont have a part number but i believe they were made by humphrey.

Good points on the software though. It would be straightforward to write a burst carb routine for a selectable beer vessel (full corny keg) that would carb quickly with little risk of overcarb.

I've never had an issue with Micromatic regulators. I am enjoying this discussion though.

I didn't want to include this in the novel I posted earlier, but it occurred to me that the carb level during bursting could be sensed by periodically shutting off input until the pressure equalized; or, even more quickly and fancily, by monitoring the decay time (although that would require a lot of empirical data).