I’m not sure exactly what you’re doing in your spreadsheet, but I certainly do find it suspicious that you wind up with a temperature of over 1000R (550F). I believe that this temperature would cause plastic tubing (and maybe tanks) to soften and rupture. This may be correct for an adiabatic compression of air, but there is certainly some temperature equalization taking place.
I did the calculations based on constant temperature (not right, but probably closer), and came up with 2 minutes and 9 seconds for 0-120 on a 1 gallon tank. This is still shorter than Viair’s 3 minutes, but a good bit closer. There does seem to be a disagreement between the “performance” and “fill rate” data.
Pneumatics-1.xlsx (182 KB)
Pneumatics-1.xlsx (182 KB)
Yup there’s a problem there. Didn’t even notice it. The compressor gets hot, but not that hot.
My assumption is isentropic compression, but I obviously have a mistake in my math somewhere. That’s what happens when you take someone who was once pretty good at aerodynamics and thermo and give them a job dealing with spacecraft…
*Is the compressor motor startup current at 100psi specified anywhere? Has anyone measured it with a current probe and a scope?
By start up, do you mean a transient? If not, it’s given on the VIAIR web page referenced in the spreadsheet: http://www.viaircorp.com/90C.html#tabs-2 as being 9A. I have not measured this myself. As my team is not using pneumatics this year, it’s unlikely that I’ll have an opportunity until after build season is over.
I don’t believe that the air compressor would be well modeled as an isentropic (reversible) process. Even if the mechanical compression were isentropic, the presence of a heat sink at the compressor output radiating/convecting heat from the compressed air into the environment would make the process irreversible.
Don’t forget that any inefficiencies are going to be released as heat. That would include inefficiencies in the mechanics of the compressor as well as the motor inside of it. That being said, tanks aren’t perfect insulators either, so some heat will escape through the walls (plastic is better than metal in that respect), so even if you model the compressor perfectly, you still have to model the whole system to get a decent model. For general FRC use, who really cares? Just use the manufactures’ specs, but as an academic exercise, there are just far too many unknowns to take into account. There are things like back pressures that get involved with fittings, pressure loss in pneumatic lines, as well as heat loss in some of the larger metallic things like the pressure switch and any manifolds that are used. All that can add up. At any rate, I think isentropic is generally a good model for a compressor you don’t know much about in terms of construction as long as you take efficiency into account. That being said, I do acknowledge that there probably is a math mistake in my spreadsheet somewhere because that temperature is extremely high (though the temperature does get extremely hot). We have melted high temp pressure tube coming off the compressor before switching to copper.
Good idea, Russ. I will try that tonight and report results. My team has three types of compressors on hand so I’ll attempt to measure starting surge current for each.
-------edit after testing--------
We tested three compressors, using an oscilloscope and current probe to measure (1) initial peak surge current, (2) average current draw during charging, and (3) time to full charge. The test system capacity was 48 cubic inch (3x Clippard AVT 32-16). In each case the duration of the initial surge current was ~0.05 second, rising to its peak in about 0.005 second and decaying exponentially to steady value after that. I have attached a typical oscilloscope plot of the compressor current, 50 seconds full horizontal scale and +/- 40 Ampere full vertical scale. This plot is for the Thomas compressor; other plots are similar in form but differ in their initial peak surge, average current draw, and time to full charge as summarized below.
Viair 90C: initial peak surge 18 Ampere, average current draw during charging 12 Ampere, time to full charge 40 seconds.
Thomas 405ADC38/12: initial peak surge 32 Ampere, average current draw during charging 10 Ampere, time to full charge 36 seconds.
Viair 250C-IG: initial peak surge 40 Ampere, average current draw during charging 9 Ampere, time to full charge 33 seconds.
Interesting that the efficiency correlates so well with initial peak surge. Assuming the voltage was constant at 12.5V throughout, the 90C needed 6.0 kJ, the Thomas 4.5 kJ, and the 250C-IG only needed 3.7 kJ from the battery to do the same job.
We’ve run similar testing and found a wide variation sometimes between compressors of identical models.
For instance, a ViAir 90c that pulled ~6amps and a Thomas that pulled under 10amps while another Thomas pulled ~12 amps.
I’d put the Thomas variations down to age and abuse, but I was surprised to see your ViAir 90C pulling so many more amps than some of ours.
How was the test system controlled? I’m curious what difference (if any) we would see controlling the compressor via a spike versus the new PCM. I’m particularly interested in the surge current for the PCM… Can we measure it both between the PCM and the compressor, and between the PCM and the PDP? The results of such a test may impact whether teams hook their PCM up through the dedicated port with a fuse, or go through a WAGO and a circuit breaker.
I am interested in that, also.
My test results above were made using the highly illegal method of series connecting a pressure switch to the compressor – the opposite of soft-starting!
Can someone try the tests Jon suggests using the PCM? We have already returned our beta hardware, and it may be a couple of days before we are ready to test this on our 2015 robot.
By time to full charge do you mean starting at 0 psi? The original question on start up current was asking what it was with 100psi already in the tank. Personallly I’d like to see the start up current and time to fill using the on-off range of the legal pressure switch. Time to fill from 0psi and the start up current at that pressure is less critical since you can fill before the match starts and should only occur once per match while the 85-115psi event can potentially happen multiple times in a match. If the fuse were to blow before a match that could easily be corrected before going on the field.
Yes, we started with the tanks empty. I was looking mostly at peak surge current with the 'scope. A stopwatch would suffice to measure refill time, and that should be much shorter than the 0-120 fill times reported above.
Also, I would expect bigger differences in the compressor comparisons for 100-120 refill times, based on their published flow curves.
BTW, the peak surge transient (50 msec) we observed would not trip a breaker or blow a fuse.
I’m no expert, but I’ve been repeating the “conventional wisdom” for years as an inspector and making teams replace their spike fuse with a breaker for the compressor. I had heard that the fuse will wear out over time due to the start up surge until it eventually pops…
Can anyone with a better understanding correlate the readings posted in this thread with the compressor breaker rule we’ve all been following for years?
I am not an expert, but my team has made some tests which may shed some light. See attached summary of our set up and results for recharging a 48 cubic inch system using three FRC-legal compressors.
When recharging from ~100 to ~120 PSI, the compressor must start against load. As Mr. V pointed out above, this case is much more pertinent to actual FRC operation than the tests I reported earlier.
The initial peak surge current observed using Viair compressors is relatively short duration (<0.1 sec), but the Thomas compressor’s initial peak surge current is much longer duration (~0.5 sec). Comparing the Thomas surge against typical automotive fuse curves suggests an explanation for the “conventional wisdom” Jon mentioned above. The Thomas compressor has much less margin against blowing a 20A fuse during surge, and historically that is the compressor on which our FRC “conventional wisdom” is based.
test set up 48.pdf (447 KB)
test set up 48.pdf (447 KB)
Excelent and very useful data, thanks very much.
What people need to keep in mind is that a fuse and the type of circuit breakers that we use are thermal devices. When they reach a certain temp they will open. With a large overload it will open almost instantaneously. With a slight overload it will open but in a longer period. With intermittent medium overloads it can eventual heat to the point where it opens.
So the old Thomas compressor if it starts multiple times during a match could certainly eventually heat up the fuse to the point it opens. I’m not sure why they choose the ATM form factor instead of the ATO/ATC form factor as used for the 20/30 amp circuits on the PDP, PDB and Spike. The ATM is not that much smaller than the ATO/ATC.
Richard HI
I try the test. noting move may be because of the spike, I use defrent spike , nothing move.
I saw your questions in the other thread. Have you been able to determine if your Thomas compressor is still functioning normally? Do you have another compressor to try?
YesI have ,
for testing Iconnect dirict itis working.
My quation i you knowhow to use the spike as relay. I use Jumber, the output 0 [V]
Ok, now I understand your question. I caused confusion by showing the control connection between the pressure switch and the Spike incorrectly in the diagram attached in my earlier post. The connection used on my team’s bench test board is shown in the attached (revA) version of that diagram.
test set up 48 revA.pdf (447 KB)
test set up 48 revA.pdf (447 KB)